@inproceedings{zheng-etal-2026-stealthgraph,
    title = "{S}tealth{G}raph: Exposing Domain-Specific Risks in {LLM}s through Knowledge-Graph-Guided Harmful Prompt Generation",
    author = "Zheng, Huawei  and
      Jiang, Xinqi  and
      Yang, Sen  and
      Ji, Shouling  and
      Wu, Yingcai  and
      Deng, Dazhen",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/check-for-anonymous-pdfs/2026.acl-long.295/",
    pages = "6519--6539",
    ISBN = "979-8-89176-390-6",
    abstract = "Large language models (LLMs) are increasingly applied in specialized domains such as finance and healthcare, where they introduce unique safety risks. Domain-specific datasets of harmful prompts remain scarce and still largely rely on manual construction; public datasets mainly focus on explicit harmful prompts, which modern LLM defenses can often detect and refuse. In contrast, implicit harmful prompts{---}expressed through indirect domain knowledge{---}are harder to detect and better reflect real-world threats. We identify two challenges: transforming domain knowledge into actionable constraints and increasing the implicitness of generated harmful prompts. To address them, we propose an end-to-end framework that first performs knowledge-graph-guided harmful prompt generation to systematically produce domain-relevant prompts, and then applies two-strategy obfuscation rewriting to convert explicit harmful prompts into implicit variants via direct and context-enhanced rewriting. This framework yields high-quality datasets combining strong domain relevance with implicitness, enabling more realistic red-teaming and advancing LLM safety research. We release our code and datasets on GitHub."
}