English
Related papers

Related papers: Guardrails as Infrastructure: Policy-First Control…

200 papers

We study how runtime enforcement against unsafe actions affects end-to-end task performance in multi-step tool using large language model (LLM) agents. Using tau-bench across Airline and Retail domains, we compare baseline Tool-Calling,…

Cryptography and Security · Computer Science 2026-03-23 Tanmay Sah , Vishal Srivastava , Dolly Sah , Kayden Jordan

Recent breakthroughs in Large Language Models (LLMs) have led to their adoption across a wide range of tasks, ranging from code generation to machine translation and sentiment analysis, etc. Red teaming/Safety alignment efforts show that…

Computation and Language · Computer Science 2024-09-25 Essa Jan , Nouar AlDahoul , Moiz Ali , Faizan Ahmad , Fareed Zaffar , Yasir Zaki

Large language models (LLMs) remain susceptible to jailbreak and direct prompt-injection attacks, yet the strongest defensive filters frequently over-refuse benign queries and degrade user experience. Previous work on jailbreak and prompt…

Computation and Language · Computer Science 2026-04-08 Purva Chiniya , Kevin Scaria , Sagar Chaturvedi

The rise of AI agent frameworks has introduced agent skills, modular packages containing instructions and executable code that dynamically extend agent capabilities. While this architecture enables powerful customization, skills execute…

Cryptography and Security · Computer Science 2026-01-16 Yi Liu , Weizhe Wang , Ruitao Feng , Yao Zhang , Guangquan Xu , Gelei Deng , Yuekang Li , Leo Zhang

Machine learning (ML) underpins foundation models in finance, healthcare, and critical infrastructure, making them targets for data poisoning, model extraction, prompt injection, automated jailbreaking, and preference-guided black-box…

Cryptography and Security · Computer Science 2025-12-30 Armstrong Foundjem , Lionel Nganyewou Tidjon , Leuson Da Silva , Foutse Khomh

Generative AI holds the promise of enabling a range of sought-after capabilities and revolutionizing workflows in various consumer and enterprise verticals. However, putting a model in production involves much more than just generating an…

LLM jailbreaks are a widespread safety challenge. Given this problem has not yet been tractable, we suggest targeting a key failure mechanism: the failure of safety to generalize across semantically equivalent inputs. We further focus the…

Guardrail models (a.k.a. safety checkers) are widely deployed to screen user inputs before they reach large language models (LLMs), serving as a primary defense against prompt injection attacks. Due to strict context constraints, these…

Cryptography and Security · Computer Science 2026-05-25 Yuanbo Zhou , Changjia Zhu , Junyu Wang , Xu He , Yan Zhai , Kun Sun , Mingkui Wei , Junjie Xiong

Large Language Models (LLMs) are prone to off-topic misuse, where users may prompt these models to perform tasks beyond their intended scope. Current guardrails, which often rely on curated examples or custom classifiers, suffer from high…

Computation and Language · Computer Science 2025-04-10 Gabriel Chua , Shing Yee Chan , Shaun Khoo

Large language model (LLM)-based AI agents extend LLM capabilities by enabling access to tools such as data sources, APIs, search engines, code sandboxes, and even other agents. While this empowers agents to perform complex tasks, LLMs may…

Software Engineering · Computer Science 2026-01-14 Aarya Doshi , Yining Hong , Congying Xu , Eunsuk Kang , Alexandros Kapravelos , Christian Kästner

Large language models (LLMs) are increasingly deployed as agents with access to executable tools, enabling direct interaction with external systems. However, most safety evaluations remain text-centric and assume that compliant language…

Software Engineering · Computer Science 2026-03-24 Shasha Yu , Fiona Carroll , Barry L. Bentley

We study the problem of policy repair for learning-based control policies in safety-critical settings. We consider an architecture where a high-performance learning-based control policy (e.g. one trained as a neural network) is paired with…

Artificial Intelligence · Computer Science 2020-08-19 Weichao Zhou , Ruihan Gao , BaekGyu Kim , Eunsuk Kang , Wenchao Li

Tool-augmented LLM agents increasingly rely on multi-step, multi-tool workflows to complete real tasks. This design expands the attack surface, because data produced by one tool can be persisted and later reused as input to another tool,…

Software Engineering · Computer Science 2026-03-16 Jiangrong Wu , Zitong Yao , Yuhong Nan , Zibin Zheng

Tool using agents often fail for operational reasons even when language understanding is strong. Common causes include invalid arguments, interface drift, weak recovery, and inefficient retry behavior. We introduce ToolMisuseBench, an…

Software Engineering · Computer Science 2026-04-03 Akshey Sigdel , Rista Baral

Indirect prompt injection attacks threaten AI agents that execute consequential actions, motivating deterministic system-level defenses. Such defenses can provably block unsafe actions by enforcing confidentiality and integrity policies,…

Cryptography and Security · Computer Science 2026-02-13 Aashish Kolluri , Rishi Sharma , Manuel Costa , Boris Köpf , Tobias Nießen , Mark Russinovich , Shruti Tople , Santiago Zanella-Béguelin

Modern AI agents execute real-world side effects through tool calls such as file operations, shell commands, HTTP requests, and database queries. A single unsafe action, including accidental deletion, credential exposure, or data…

Artificial Intelligence · Computer Science 2026-05-07 Chenglin Yang

Agent Skills, structured packages of procedural knowledge loaded into an LLM agent at inference time, are widely reported to improve task pass rates by an average of 16.2~percentage points across diverse domains. Yet the same benchmarks…

Artificial Intelligence · Computer Science 2026-05-26 Samuel Jacob Chacko , James Hugglestone , Chashi Mahiul Islam , Xiuwen Liu

In this study, we disclose a worrying new vulnerability in Large Language Models (LLMs), which we term \textbf{involuntary jailbreak}. Unlike existing jailbreak attacks, this weakness is distinct in that it does not involve a specific…

Cryptography and Security · Computer Science 2025-12-30 Yangyang Guo , Yangyan Li , Mohan Kankanhalli

The integration of Large Language Models (LLMs) into software engineering has revolutionized code generation, enabling unprecedented productivity through promptware and autonomous AI agents. However, this transformation introduces…

Software Engineering · Computer Science 2025-08-19 Satyam Kumar Navneet , Joydeep Chandra

The integration of tool use into large language models (LLMs) enables agentic systems with real-world impact. In the meantime, unlike standalone LLMs, compromised agents can execute malicious workflows with more consequential impact,…

Cryptography and Security · Computer Science 2025-02-17 Jizhou Chen , Samuel Lee Cong