Reproduction Package - Fixing, Breaking, or Faking It? An Execution-Calibrated Evaluation of LLM Vulnerability Patching and the Limits of LLM-as-Judge

Replication package Fixing, Breaking, or Faking It? An Execution-Calibrated Evaluation of LLM Vulnerability Patching and the Limits of LLM-as-Judge Patrick Deininger1,2 and Wolfgang Slany1 1 Institute of Software Engineering and AI, Graz University of Technology 2 Institute of Computer Science and AI, FH JOANNEUM Submitted to Journal of Cybersecurity and Privacy (MDPI). Archive DOI: https://doi.org/10.5281/zenodo.21645415 · Licence: CC BY 4.0 (see LICENSE)