ALIBI: Adaptive Agentic Attacks on LLM-Based Vulnerability Detectors via Adversarial Code Comments

Abstract

Large language models are increasingly deployed for security-sensitive tasks such as vulnerability detection and code review. Their reliance on natural-language context embedded in source code exposes a previously underexplored attack surface: adversarial comments that can influence a detector's reasoning without changing program behavior. We study LLM-based vulnerability detectors against a new adversary: a coding agent that implements new functionality, deliberately introduces vulnerabilities,

Categories

@misc{wu2026alibi,
  title = {{ALIBI: Adaptive Agentic Attacks on LLM-Based Vulnerability Detectors via Adversarial Code Comments}},
  author = {Zixuan Wu and Cristina Nita-Rotaru},
  year = {2026},
  month = jul,
  eprint = {2607.24964},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2607.24964}
}