Leveraging Large Language Models to Detect npm Malicious Packages

2024-03-18

AI Summary

High false-positive rates and substantial manual analysis costs hinder effective detection of malicious packages in the npm ecosystem. Method: We propose a hybrid automated detection framework integrating lightweight static analysis with large language models (LLMs). First, CodeQL—guided by 39 domain-specific rules—performs coarse-grained pre-screening, eliminating 77.9% of non-suspicious files. Subsequently, GPT-4 conducts fine-grained semantic analysis on the remaining code snippets. Contribution/Results: This is the first systematic evaluation of GPT-3/GPT-4 for JavaScript-based malicious package detection and the first end-to-end LLM-driven security analysis workflow for npm. Evaluated on a benchmark of 5,115 npm packages, our approach achieves 99% precision and 97% F1-score—outperforming pure CodeQL by +16% precision and +9% F1—while reducing LLM inference cost by over 60%.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Existing malicious code detection techniques demand the integration of multiple tools to detect different malware patterns, often suffering from high misclassification rates. Therefore, malicious code detection techniques could be enhanced by adopting advanced, more automated approaches to achieve high accuracy and a low misclassification rate. The goal of this study is to aid security analysts in detecting malicious packages by empirically studying the effectiveness of Large Language Models (LLMs) in detecting malicious code. We present SocketAI, a malicious code review workflow to detect malicious code. To evaluate the effectiveness of SocketAI, we leverage a benchmark dataset of 5,115 npm packages, of which 2,180 packages have malicious code. We conducted a baseline comparison of GPT-3 and GPT-4 models with the state-of-the-art CodeQL static analysis tool, using 39 custom CodeQL rules developed in prior research to detect malicious Javascript code. We also compare the effectiveness of static analysis as a pre-screener with SocketAI workflow, measuring the number of files that need to be analyzed. and the associated costs. Additionally, we performed a qualitative study to understand the types of malicious activities detected or missed by our workflow. Our baseline comparison demonstrates a 16% and 9% improvement over static analysis in precision and F1 scores, respectively. GPT-4 achieves higher accuracy with 99% precision and 97% F1 scores, while GPT-3 offers a more cost-effective balance at 91% precision and 94% F1 scores. Pre-screening files with a static analyzer reduces the number of files requiring LLM analysis by 77.9% and decreases costs by 60.9% for GPT-3 and 76.1% for GPT-4. Our qualitative analysis identified data theft, execution of arbitrary code, and suspicious domain categories as the top detected malicious packages.

Problem

• Malware Detection
• npm Packages
• Automated Identification

Innovation

• SocketAI
• Large Language Models
• malware detection