Helpful or Harmful? Evaluating LLM-Assisted Vulnerability Patching Via a Human Study

Software vulnerability remediation is a cognitively demanding task that requires specialized security expertise often lacking in general developers. In the meantime, Large Language Models LLMs assisted tools show potential in vulnerability detection, location, and repair tasks. Hypothesis: While...