Shifting the Lens: Detecting Malicious npm Packages using Large Language Models

• URL: http://arxiv.org/abs/2403.12196v2
• Date: Fri, 9 Aug 2024 16:29:44 GMT
• Title: Shifting the Lens: Detecting Malicious npm Packages using Large Language Models
• Authors: Nusrat Zahan, Philipp Burckhardt, Mikola Lysenko, Feross Aboukhadijeh, Laurie Williams,
• Abstract summary: Existing malicious code detection techniques often suffer from high misclassification rates. We present SecurityAI, a malicious code review workflow to detect malicious code using ChatGPT. Our baseline comparison demonstrates a 16% and 9% improvement over static analysis in precision and F1 scores.
• Score: 4.479741014073169
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Existing malicious code detection techniques can aid the manual review process by predicting which packages are likely to be malicious. However, these techniques often suffer 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 assist security analysts in detecting malicious packages through the empirical study of using Large Language Models (LLMs) to detect malicious code in the npm ecosystem. We present SecurityAI, a malicious code review workflow to detect malicious code using ChatGPT. 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 compare the effectiveness of static analysis as a pre-screener with SecurityAI 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 packages detected or missed by our workflow. Our baseline comparison demonstrates a 16% and 9% improvement over static analysis in precision and F1 scores, respectively. We attained precision and F1 scores of 91% and 94% for GPT-3, and 99% & 97% for GPT-4, respectively, with GPT-3 offering a cost-effective balance. 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, hidden backdoors, and suspicious domain connection categories as the top detected malicious packages.

Related papers

• Preference Optimization for Reasoning with Pseudo Feedback [100.62603571434167]
  We introduce a novel approach to generate pseudo feedback for reasoning tasks by framing the labeling of solutions as an evaluation against associated test cases. We conduct experiments on both mathematical reasoning and coding tasks using pseudo feedback for preference optimization, and observe improvements across both tasks.
  arXiv  Detail & Related papers  (2024-11-25T12:44:02Z)
• Automated Software Vulnerability Static Code Analysis Using Generative Pre-Trained Transformer Models [0.8192907805418583]
  Generative Pre-Trained Transformer models have been shown to be surprisingly effective at a variety of natural language processing tasks. We evaluate the effectiveness of open source GPT models for the task of automatic identification of the presence of vulnerable code syntax.
  arXiv  Detail & Related papers  (2024-07-31T23:33:26Z)
• Networks of Networks: Complexity Class Principles Applied to Compound AI Systems Design [63.24275274981911]
  Compound AI Systems consisting of many language model inference calls are increasingly employed. In this work, we construct systems, which we call Networks of Networks (NoNs) organized around the distinction between generating a proposed answer and verifying its correctness. We introduce a verifier-based judge NoN with K generators, an instantiation of "best-of-K" or "judge-based" compound AI systems.
  arXiv  Detail & Related papers  (2024-07-23T20:40:37Z)
• SORRY-Bench: Systematically Evaluating Large Language Model Safety Refusal Behaviors [64.9938658716425]
  Existing evaluations of large language models' (LLMs) ability to recognize and reject unsafe user requests face three limitations. First, existing methods often use coarse-grained of unsafe topics, and are over-representing some fine-grained topics. Second, linguistic characteristics and formatting of prompts are often overlooked, like different languages, dialects, and more -- which are only implicitly considered in many evaluations. Third, existing evaluations rely on large LLMs for evaluation, which can be expensive.
  arXiv  Detail & Related papers  (2024-06-20T17:56:07Z)
• Do Neutral Prompts Produce Insecure Code? FormAI-v2 Dataset: Labelling Vulnerabilities in Code Generated by Large Language Models [3.4887856546295333]
  This study provides a comparative analysis of state-of-the-art large language models (LLMs) We analyze how likely they generate vulnerabilities when writing simple C programs using a neutral zero-shot prompt.
  arXiv  Detail & Related papers  (2024-04-29T01:24:14Z)
• Data-Free Hard-Label Robustness Stealing Attack [67.41281050467889]
  We introduce a novel Data-Free Hard-Label Robustness Stealing (DFHL-RS) attack in this paper. It enables the stealing of both model accuracy and robustness by simply querying hard labels of the target model. Our method achieves a clean accuracy of 77.86% and a robust accuracy of 39.51% against AutoAttack.
  arXiv  Detail & Related papers  (2023-12-10T16:14:02Z)
• Zero-Shot Detection of Machine-Generated Codes [83.0342513054389]
  This work proposes a training-free approach for the detection of LLMs-generated codes. We find that existing training-based or zero-shot text detectors are ineffective in detecting code. Our method exhibits robustness against revision attacks and generalizes well to Java codes.
  arXiv  Detail & Related papers  (2023-10-08T10:08:21Z)
• Can Large Language Models Find And Fix Vulnerable Software? [0.0]
  GPT-4 identified approximately four times the vulnerabilities than its counterparts. It provided viable fixes for each vulnerability, demonstrating a low rate of false positives. GPT-4's code corrections led to a 90% reduction in vulnerabilities, requiring only an 11% increase in code lines.
  arXiv  Detail & Related papers  (2023-08-20T19:33:12Z)
• G$^2$uardFL: Safeguarding Federated Learning Against Backdoor Attacks through Attributed Client Graph Clustering [116.4277292854053]
  Federated Learning (FL) offers collaborative model training without data sharing. FL is vulnerable to backdoor attacks, where poisoned model weights lead to compromised system integrity. We present G$2$uardFL, a protective framework that reinterprets the identification of malicious clients as an attributed graph clustering problem.
  arXiv  Detail & Related papers  (2023-06-08T07:15:04Z)
• Dataflow Analysis-Inspired Deep Learning for Efficient Vulnerability Detection [17.761541379830373]
  DeepDFA is a dataflow analysis-inspired graph learning framework. It was trained in 9 minutes, 75x faster than the highest-performing baseline model. It detected 8.7 out of 17 vulnerabilities on average across folds and was able to distinguish between patched and buggy versions.
  arXiv  Detail & Related papers  (2022-12-15T19:49:27Z)
• Malicious Code Detection: Run Trace Output Analysis by LSTM [0.0]
  We propose a methodological framework for detecting malicious code by analyzing run trace outputs by Long Short-Term Memory (LSTM) We created our dataset from run trace outputs obtained from dynamic analysis of PE files. Experiments showed that the ISM achieved an accuracy of 87.51% and a false positive rate of 18.34%, while BSM achieved an accuracy of 99.26% and a false positive rate of 2.62%.
  arXiv  Detail & Related papers  (2021-01-14T15:00:42Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.