English
Related papers

Related papers: Auditing Agent Harness Safety

200 papers

This position paper argues that, for long-horizon tasks evaluated across models with comparable frontier capability, the agent execution harness, namely the infrastructure layer that governs context construction, tool interaction,…

Artificial Intelligence · Computer Science 2026-05-26 Yunbei Zhang , Janet Wang , Yingqiang Ge , Weijie Xu , Jihun Hamm , Chandan K. Reddy

LLM agents increasingly perform end-to-end ML engineering tasks where success is judged by a single scalar test metric. This creates a structural vulnerability: an agent can increase the reported score by compromising the evaluation…

Artificial Intelligence · Computer Science 2026-03-13 Yonas Atinafu , Robin Cohen

Personal AI agents like OpenClaw run with elevated privileges on users' local machines, where a single successful prompt injection can leak credentials, redirect financial transactions, or destroy files. This threat goes well beyond…

Artificial Intelligence · Computer Science 2026-04-07 Bowen Wei , Yunbei Zhang , Jinhao Pan , Kai Mei , Xiao Wang , Jihun Hamm , Ziwei Zhu , Yingqiang Ge

Large vision-language model (LVLM)-based web agents are emerging as powerful tools for automating complex online tasks. However, when deployed in real-world environments, they face serious security risks, motivating the design of security…

Cryptography and Security · Computer Science 2026-04-15 Zonghao Ying , Yangguang Shao , Jianle Gan , Gan Xu , Wenxin Zhang , Quanchen Zou , Junzheng Shi , Zhenfei Yin , Mingchuan Zhang , Aishan Liu , Xianglong Liu

We read twelve well-known LLM agent benchmark papers and recorded, dimension by dimension, what each paper actually says about how its evaluation was run. The motivation came from a familiar frustration: two papers will report results on…

Machine Learning · Computer Science 2026-05-21 Mahdi Naser Moghadasi , Faezeh Ghaderi

AI agents have been developed for complex real-world tasks from coding to customer service. But AI agent evaluations suffer from many challenges that undermine our understanding of how well agents really work. We introduce the Holistic…

Standard benchmarks fixate on how well large language model (LLM) agents perform in finance, yet say little about whether they are safe to deploy. We argue that accuracy metrics and return-based scores provide an illusion of reliability,…

General Finance · Quantitative Finance 2025-06-03 Zichen Chen , Jiaao Chen , Jianda Chen , Misha Sra

Large language models (LLMs) are increasingly deployed as agents in various contexts by providing tools at their disposal. However, LLM agents can exhibit unpredictable behaviors, including taking undesirable and/or unsafe actions. In order…

Artificial Intelligence · Computer Science 2025-10-27 Ilija Lichkovski , Alexander Müller , Mariam Ibrahim , Tiwai Mhundwa

Autonomous agents powered by large language models (LLMs) are increasingly deployed in real-world applications requiring complex, long-horizon workflows. However, existing benchmarks predominantly focus on atomic tasks that are…

Computation and Language · Computer Science 2025-08-13 Weixuan Wang , Dongge Han , Daniel Madrigal Diaz , Jin Xu , Victor Rühle , Saravan Rajmohan

LLM agents call tools, query databases, delegate tasks, and trigger external side effects. Once an agent system can act in the world, the question is no longer only whether harmful actions can be prevented--it is whether those actions…

Artificial Intelligence · Computer Science 2026-04-08 Yi Nian , Aojie Yuan , Haiyue Zhang , Jiate Li , Yue Zhao

Language model (LM) agents have demonstrated significant potential for automating real-world tasks, yet they pose a diverse array of potential, severe risks in safety-critical scenarios. In this work, we identify a significant gap between…

Artificial Intelligence · Computer Science 2025-08-20 Yuzhi Tang , Tianxiao Li , Elizabeth Li , Chris J. Maddison , Honghua Dong , Yangjun Ruan

Agentic AI systems -- Large Language Models (LLMs) augmented with planning, tool use, memory, and long-horizon interactions -- can execute complex tasks autonomously, but their multi-step trajectories introduce new failure modes that…

Artificial Intelligence · Computer Science 2026-05-26 Jinhu Qi , Muzhi Li , Jiahong Liu , Yuqin Shu , Dianzhi Yu , Shicheng Ma , Wenqian Cui , Yiyang Zhao , Yiyi Chen , Ruoxi Jiang , Irwin King , Zenglin Xu

Operational safety at mission-critical work sites is a top priority given the complex and hazardous nature of daily tasks. This paper presents the Human-Agent Risk Navigation and Event Safety System (HARNESS), a modular AI framework…

The rapid evolution of Large Multimodal Models (LMMs) has enabled agents to perform complex digital and physical tasks, yet their deployment as autonomous decision-makers introduces substantial unintentional behavioral safety risks.…

Artificial Intelligence · Computer Science 2026-03-30 Yuxuan Li , Yi Lin , Peng Wang , Shiming Liu , Xuetao Wei

Large language model (LLM) agents are increasingly deployed to automate productivity tasks (e.g., email, scheduling, document management), but evaluating them on live services is risky due to potentially irreversible changes. Existing…

Large Language Models (LLMs) exhibit substantial promise in enhancing task-planning capabilities within embodied agents due to their advanced reasoning and comprehension. However, the systemic safety of these agents remains an underexplored…

Artificial Intelligence · Computer Science 2025-04-22 Yuting Huang , Leilei Ding , Zhipeng Tang , Tianfu Wang , Xinrui Lin , Wuyang Zhang , Mingxiao Ma , Yanyong Zhang

Large language model (LLM)-based techniques have achieved notable progress in generating harnesses for program fuzzing. However, applying them to arbitrary functions (especially internal functions) \textit{at scale} remains challenging due…

Cryptography and Security · Computer Science 2025-12-12 Kang Yang , Yunhang Zhang , Zichuan Li , Guanhong Tao , Jun Xu , Xiaojing Liao

The advances made by Large Language Models (LLMs) have led to the pursuit of LLM agents that can solve intricate, multi-step reasoning tasks. As with any research pursuit, benchmarking and evaluation are key corner stones to efficient and…

Artificial Intelligence · Computer Science 2024-04-10 Luca Gioacchini , Giuseppe Siracusano , Davide Sanvito , Kiril Gashteovski , David Friede , Roberto Bifulco , Carolin Lawrence

Multi-agent Large Language Model (LLM) systems create privacy risks that current benchmarks cannot measure. When agents coordinate on tasks, sensitive data passes through inter-agent messages, shared memory, and tool arguments, all pathways…

Artificial Intelligence · Computer Science 2026-03-31 Faouzi El Yagoubi , Godwin Badu-Marfo , Ranwa Al Mallah

We introduce WorkBench: a benchmark dataset for evaluating agents' ability to execute tasks in a workplace setting. WorkBench contains a sandbox environment with five databases, 26 tools, and 690 tasks. These tasks represent common business…

Computation and Language · Computer Science 2024-08-06 Olly Styles , Sam Miller , Patricio Cerda-Mardini , Tanaya Guha , Victor Sanchez , Bertie Vidgen