arXiv subjects

Limin Jia

Publications and source records attributed to Limin Jia.

16 recordsLinked to original sources

arXiv2026

Taint-Style Vulnerability Detection and Confirmation for Node.js Packages Using LLM Agent Reasoning

The rapidly evolving Node$.$js ecosystem currently includes millions of packages and is a critical part of modern software supply chains, making vulnerability detection of Node$.$js packages increasingly important. However, traditional program analysis struggles in this setting because of dynamic JavaScript features and the large number of package dependencies. Recent advances in large language models (LLMs) and the emerging paradigm of LLM-based agents offer an alternative to handcrafted program models. This raises the question of whether an LLM-centric, tool-augmented approach can effectively detect and confirm taint-style vulnerabilities (e.g., arbitrary command injection) in Node$.$js packages. We implement LLMVD$.$js, a multi-stage agent pipeline to scan code, propose vulnerabilities, generate proof-of-concept exploits, and validate them through lightweight execution oracles; and systematically evaluate its effectiveness in taint-style vulnerability detection and confirmation in Node$.$js packages without dedicated static/dynamic analysis engines for path derivation. For packages from public benchmarks, LLMVD$.$js confirms 84% of the vulnerabilities, compared to less than 22% for prior program analysis tools. It also outperforms a prior LLM-program-analysis hybrid approach while requiring neither vulnerability annotations nor prior vulnerability reports. When evaluated on a set of 260 recently released packages (without vulnerability groundtruth information), traditional tools produce validated exploits for few ($\leq 2$) packages, while LLMVD$.$js generates validated exploits for 36 packages.

cs.CR↗

arXiv2025

On the Difficulty of Selecting Few-Shot Examples for Effective LLM-based Vulnerability Detection

Large language models (LLMs) have demonstrated impressive capabilities across a wide range of coding tasks, including summarization, translation, completion, and code generation. Despite these advances, detecting code vulnerabilities remains a challenging problem for LLMs. In-context learning (ICL) has emerged as an effective mechanism for improving model performance by providing a small number of labeled examples within the prompt. Prior work has shown, however, that the effectiveness of ICL depends critically on how these few-shot examples are selected. In this paper, we study two intuitive criteria for selecting few-shot examples for ICL in the context of code vulnerability detection. The first criterion leverages model behavior by prioritizing samples on which the LLM consistently makes mistakes, motivated by the intuition that such samples can expose and correct systematic model weaknesses. The second criterion selects examples based on semantic similarity to the query program, using k-nearest-neighbor retrieval to identify relevant contexts. We conduct extensive evaluations using open-source LLMs and datasets spanning multiple programming languages. Our results show that for Python and JavaScript, careful selection of few-shot examples can lead to measurable performance improvements in vulnerability detection. In contrast, for C and C++ programs, few-shot example selection has limited impact, suggesting that more powerful but also more expensive approaches, such as re-training or fine-tuning, may be required to substantially improve model performance.

cs.SE↗

arXiv2025

Learning to Triage Vulnerability Reports from Program Analysis: An Empirical Study in Node.js

Program analysis tools often produce large volumes of candidate vulnerability reports that require costly manual review, creating a practical challenge: how can security analysts prioritize the reports most likely to be true vulnerabilities? This paper investigates whether machine learning can be applied to prioritizing vulnerabilities reported by program analysis tools. We focus on Node$.$js packages and collect a benchmark of 1,883 Node$.$js packages, each containing one reported ACE or ACI vulnerability. We evaluate a variety of machine learning approaches, including classical models, graph neural networks (GNNs), large language models (LLMs), and hybrid models that combine GNNs and LLMs, trained on data derived from program analysis tool outputs (NodeMedic-FINE in our case study). The top LLM achieves $F_{1}{=}0.915$, while the best provenance-graph-based method achieves $F_{1}{=}0.904$. On the reports that the upstream tool flags but cannot automatically confirm, at a target of recovering 90% of the exploitable reports, the leading model eliminates 75% of the benign reports from manual review. If the best model is tuned to operate at a precision level of 0.8 (i.e., allowing 20% false positives among all warnings), our approach can report 99.2% of exploitable taint flows while missing only 0.8%, demonstrating strong potential for real-world vulnerability triage.

cs.CR↗

arXiv2025

WAMI: Compilation to WebAssembly through MLIR without Losing Abstraction

WebAssembly (Wasm) is a portable bytecode format that serves as a compilation target for high-level languages, enabling their secure and efficient execution across diverse platforms, including web browsers and embedded systems. To improve support for high-level languages without incurring significant code size or performance overheads, Wasm continuously evolves by integrating high-level features such as Garbage Collection and Stack Switching. However, existing compilation approaches either lack reusable design -- requiring redundant implementation efforts for each language -- or lose abstraction by lowering high-level constructs into low-level shared representations like LLVM IR, which hinder the adoption of high-level features. MLIR compiler infrastructure provides the compilation pipeline with multiple levels of abstraction, preserving high-level abstractions throughout the compilation pipeline, yet the current MLIR pipeline relies on the LLVM backend for Wasm code generation, thereby inheriting LLVM's limitations. This paper presents a novel compilation pipeline for Wasm, featuring Wasm dialects explicitly designed to represent high-level Wasm constructs within MLIR. Our approach enables direct generation of high-level Wasm code from corresponding high-level MLIR dialects without losing abstraction, providing a modular and extensible way to incorporate high-level Wasm features. We illustrate this extensibility through a case study that leverages Stack Switching, a recently introduced high-level feature of Wasm. Performance evaluations on PolyBench benchmarks show that our pipeline, benefiting from optimizations within the MLIR and Wasm ecosystems, produces code with at most 7.7\% slower, and faster in some execution environments, compared to LLVM-based compilers.

cs.PL↗

arXiv2023

Transfer Attacks and Defenses for Large Language Models on Coding Tasks

Modern large language models (LLMs), such as ChatGPT, have demonstrated impressive capabilities for coding tasks including writing and reasoning about code. They improve upon previous neural network models of code, such as code2seq or seq2seq, that already demonstrated competitive results when performing tasks such as code summarization and identifying code vulnerabilities. However, these previous code models were shown vulnerable to adversarial examples, i.e. small syntactic perturbations that do not change the program's semantics, such as the inclusion of "dead code" through false conditions or the addition of inconsequential print statements, designed to "fool" the models. LLMs can also be vulnerable to the same adversarial perturbations but a detailed study on this concern has been lacking so far. In this paper we aim to investigate the effect of adversarial perturbations on coding tasks with LLMs. In particular, we study the transferability of adversarial examples, generated through white-box attacks on smaller code models, to LLMs. Furthermore, to make the LLMs more robust against such adversaries without incurring the cost of retraining, we propose prompt-based defenses that involve modifying the prompt to include additional information such as examples of adversarially perturbed code and explicit instructions for reversing adversarial perturbations. Our experiments show that adversarial examples obtained with a smaller code model are indeed transferable, weakening the LLMs' performance. The proposed defenses show promise in improving the model's resilience, paving the way to more robust defensive solutions for LLMs in code-related applications.

cs.LG↗

arXiv2021

Towards a Formal Foundation of Intermittent Computing

Intermittently powered devices enable new applications in harsh or inaccessible environments, such as space or in-body implants, but also introduce problems in programmability and correctness. Researchers have developed programming models to ensure that programs make progress and do not produce erroneous results due to memory inconsistencies caused by intermittent executions. As the technology has matured, more and more features are added to intermittently powered devices, such as I/O. Prior work has shown that all existing intermittent execution models have problems with repeated device or sensor inputs (RIO). RIOs could leave intermittent executions in an inconsistent state. Such problems and the proliferation of existing intermittent execution models necessitate a formal foundation for intermittent computing. In this paper, we formalize intermittent execution models, their correctness properties with respect to memory consistency and inputs, and identify the invariants needed to prove systems correct. We prove equivalence between several existing intermittent systems. To address RIO problems, we define an algorithm for identifying variables affected by RIOs that need to be restored after reboot and prove the algorithm correct. Finally, we implement the algorithm in a novel intermittent runtime system that is correct with respect to input operations and evaluate its performance.

cs.PL↗

arXiv2021

Automatically Enforcing Fresh and Consistent Inputs in Intermittent Systems

Intermittently powered energy-harvesting devices enable new applications in inaccessible environments. Program executions must be robust to unpredictable power failures, introducing new challenges in programmability and correctness. One hard problem is that input operations have implicit constraints, embedded in the behavior of continuously powered executions, on when input values can be collected and used. This paper aims to develop a formal framework for enforcing these constraints. We identify two key properties -- freshness (i.e., uses of inputs must satisfy the same time constraints as in continuous executions) and temporal consistency (i.e., the collection of a set of inputs must satisfy the same time constraints as in continuous executions). We formalize these properties and show that they can be enforced using atomic regions. We develop Ocelot, an LLVM-based analysis and transformation tool targeting Rust, to enforce these properties automatically. Ocelot provides the programmer with annotations to express these constraints and infers atomic region placement in a program to satisfy them. We then formalize Ocelot's design and show that Ocelot generates correct programs with little performance cost or code changes.

cs.PL↗

arXiv2021

Session Logical Relations for Noninterference

Information flow control type systems statically restrict the propagation of sensitive data to ensure end-to-end confidentiality. The property to be shown is noninterference, asserting that an attacker cannot infer any secrets from made observations. Session types delimit the kinds of observations that can be made along a communication channel by imposing a protocol of message exchange. These protocols govern the exchange along a single channel and leave unconstrained the propagation along adjacent channels. This paper contributes an information flow control type system for linear session types. The type system stands in close correspondence with intuitionistic linear logic. Intuitionistic linear logic typing ensures that process configurations form a tree such that client processes are parent nodes and provider processes child nodes. To control the propagation of secret messages, the type system is enriched with secrecy levels and arranges these levels to be aligned with the configuration tree. Two levels are associated with every process: the maximal secrecy denoting the process' security clearance and the running secrecy denoting the highest level of secret information obtained so far. The computational semantics naturally stratifies process configurations such that higher-secrecy processes are parents of lower-secrecy ones, an invariant enforced by typing. Noninterference is stated in terms of a logical relation that is indexed by the secrecy-level-enriched session types. The logical relation contributes a novel development of logical relations for session typed languages as it considers open configurations, allowing for more nuanced equivalence statement.

cs.PL↗

arXiv2021

Containing Malicious Package Updates in npm with a Lightweight Permission System

The large amount of third-party packages available in fast-moving software ecosystems, such as Node.js/npm, enables attackers to compromise applications by pushing malicious updates to their package dependencies. Studying the npm repository, we observed that many packages in the npm repository that are used in Node.js applications perform only simple computations and do not need access to filesystem or network APIs. This offers the opportunity to enforce least-privilege design per package, protecting applications and package dependencies from malicious updates. We propose a lightweight permission system that protects Node.js applications by enforcing package permissions at runtime. We discuss the design space of solutions and show that our system makes a large number of packages much harder to be exploited, almost for free.

cs.CR↗

arXiv2021

First-order Gradual Information Flow Types with Gradual Guarantees

Information flow type systems enforce the security property of noninterference by detecting unauthorized data flows at compile-time. However, they require precise type annotations, making them difficult to use in practice as much of the legacy infrastructure is written in untyped or dynamically-typed languages. Gradual typing seamlessly integrates static and dynamic typing, providing the best of both approaches, and has been applied to information flow control, where information flow monitors are derived from gradual security types. Prior work on gradual information flow typing uncovered tensions between noninterference and the dynamic gradual guarantee -- the property that less precise security type annotations in a program should not cause more runtime errors. This paper re-examines the connection between gradual information flow types and information flow monitors to identify the root cause of the tension between the gradual guarantees and noninterference. We develop runtime semantics for a simple imperative language with gradual information flow types that provides both noninterference and gradual guarantees. We leverage a proof technique developed for FlowML and reduce noninterference proofs to preservation proofs.

cs.PL↗

arXiv2020

On the Generation of Disassembly Ground Truth and the Evaluation of Disassemblers

When a software transformation or software security task needs to analyze a given program binary, the first step is often disassembly. Since many modern disassemblers have become highly accurate on many binaries, we believe reliable disassembler benchmarking requires standardizing the set of binaries used and the disassembly ground truth about these binaries. This paper presents (i) a first version of our work-in-progress disassembly benchmark suite, which comprises 879 binaries from diverse projects compiled with multiple compilers and optimization settings, and (ii) a novel disassembly ground truth generator leveraging the notion of "listing files", which has broad support by Clang, GCC, ICC, and MSVC. In additional, it presents our evaluation of four prominent open-source disassemblers using this benchmark suite and a custom evaluation system. Our entire system and all generated data are maintained openly on GitHub to encourage community adoption.

cs.PL↗

arXiv2019

Uncovering Information Flow Policy Violations in C Programs

Programmers of cryptographic applications written in C need to avoid common mistakes such as sending private data over public channels, modifying trusted data with untrusted functions, or improperly ordering protocol steps. These secrecy, integrity, and sequencing policies can be cumbersome to check with existing general-purpose tools. We have developed a novel means of specifying and uncovering violations of these policies that allows for a much lighter-weight approach than previous tools. We embed the policy annotations in C's type system via a source-to-source translation and leverage existing C compilers to check for policy violations, achieving high performance and scalability. We show through case studies of recent cryptographic libraries and applications that our work is able to express detailed policies for large bodies of C code and can find subtle policy violations. To gain formal understanding of our policy annotations, we show formal connections between the policy annotations and an information flow type system and prove a noninterference guarantee.

cs.CR↗

arXiv2015

A Program Logic for Verifying Secure Routing Protocols

The Internet, as it stands today, is highly vulnerable to attacks. However, little has been done to understand and verify the formal security guarantees of proposed secure inter-domain routing protocols, such as Secure BGP (S-BGP). In this paper, we develop a sound program logic for SANDLog-a declarative specification language for secure routing protocols for verifying properties of these protocols. We prove invariant properties of SANDLog programs that run in an adversarial environment. As a step towards automated verification, we implement a verification condition generator (VCGen) to automatically extract proof obligations. VCGen is integrated into a compiler for SANDLog that can generate executable protocol implementations; and thus, both verification and empirical evaluation of secure routing protocols can be carried out in this unified framework. To validate our framework, we encoded several proposed secure routing mechanisms in SANDLog, verified variants of path authenticity properties by manually discharging the generated verification conditions in Coq, and generated executable code based on SANDLog specification and ran the code in simulation.

cs.LO↗

arXiv2015

Equivalence-based Security for Querying Encrypted Databases: Theory and Application to Privacy Policy Audits

Motivated by the problem of simultaneously preserving confidentiality and usability of data outsourced to third-party clouds, we present two different database encryption schemes that largely hide data but reveal enough information to support a wide-range of relational queries. We provide a security definition for database encryption that captures confidentiality based on a notion of equivalence of databases from the adversary's perspective. As a specific application, we adapt an existing algorithm for finding violations of privacy policies to run on logs encrypted under our schemes and observe low to moderate overheads.

cs.CR↗

arXiv2015

System M: A Program Logic for Code Sandboxing and Identification

Security-sensitive applications that execute untrusted code often check the code's integrity by comparing its syntax to a known good value or sandbox the code to contain its effects. System M is a new program logic for reasoning about such security-sensitive applications. System M extends Hoare Type Theory (HTT) to trace safety properties and, additionally, contains two new reasoning principles. First, its type system internalizes logical equality, facilitating reasoning about applications that check code integrity. Second, a confinement rule assigns an effect type to a computation based solely on knowledge of the computation's sandbox. We prove the soundness of system M relative to a step-indexed trace-based semantic model. We illustrate both new reasoning principles of system M by verifying the main integrity property of the design of Memoir, a previously proposed trusted computing system for ensuring state continuity of isolated security-sensitive applications.

cs.CR↗

arXiv2011

A Logical Method for Policy Enforcement over Evolving Audit Logs

We present an iterative algorithm for enforcing policies represented in a first-order logic, which can, in particular, express all transmission-related clauses in the HIPAA Privacy Rule. The logic has three features that raise challenges for enforcement --- uninterpreted predicates (used to model subjective concepts in privacy policies), real-time temporal properties, and quantification over infinite domains (such as the set of messages containing personal information). The algorithm operates over audit logs that are inherently incomplete and evolve over time. In each iteration, the algorithm provably checks as much of the policy as possible over the current log and outputs a residual policy that can only be checked when the log is extended with additional information. We prove correctness and termination properties of the algorithm. While these results are developed in a general form, accounting for many different sources of incompleteness in audit logs, we also prove that for the special case of logs that maintain a complete record of all relevant actions, the algorithm effectively enforces all safety and co-safety properties. The algorithm can significantly help automate enforcement of policies derived from the HIPAA Privacy Rule.

cs.LO↗