一句话：首次对 LLM 应用商店中的应用抢占与克隆进行大规模测量，并给出自动化检测工具与恶意行为比例。
移动应用生态中长期存在应用抢占与克隆等仿冒手段；随着 GPT Store、FlowGPT 等 LLM 应用商店兴起，同类威胁再度浮现，但业界缺乏自动化检测机制。本文首次对 LLM 应用抢占与克隆开展大规模测量，提出工具 LLMappCrazy：它实现 14 种抢占式名称生成技术，并融合 Levenshtein 距离与 BERT 语义分析，通过功能相似度检测克隆。基于前 1000 个应用名的变体，作者发现 5000 余个抢占式 LLM 应用，并在六大平台观察到 13,325 个克隆案例；抽样显示 4.7% 的抢占应用与 18.4% 的克隆应用存在钓鱼、恶意软件分发、虚假内容传播与广告注入等恶意行为。
Impersonation tactics, such as app squatting and app cloning, have posed longstanding challenges in the mobile app industry, where malicious actors exploit the names and reputations of popular apps to deceive users. With the rapid growth of large language model (LLM) stores like GPT Store and FlowGPT, these issues have similarly surfaced, highlighting the urgent need for robust industry standards and automated detection mechanisms to safeguard the LLM app ecosystem and protect users from fraudulent practices. In this study, we present the first large-scale analysis of LLM app squatting and cloning using our custom-built tool, LLMappCrazy. LLMappCrazy covers 14 squatting generation techniques and integrates Levenshtein distance and BERT-based semantic analysis to detect cloning by analyzing app functional similarities. Using this tool, we generated variations of the top 1000 app names and found over 5,000 squatting LLM apps in the dataset. Additionally, we observed 13,325 cloning cases across six major platforms. After sampling, we find that 4.7% of the squatting apps and 18.4% of the cloning apps exhibited malicious behavior, including phishing, malware distribution, fake content dissemination, and aggressive ad injection. Our work provides actionable insights for industry stakeholders to address these growing threats and foster a safer, more trustworthy LLM app ecosystem.