arXiv AI
Jul 22
Broken Gates: Re-evaluating Web Bot Defenses in the Age of LLM Agents
arXiv:2607. 18659v1 Announce Type: cross Abstract: LLM-based browser agents are rapidly changing the threat landscape for web security.
Broken Gates: Re-evaluating Web Bot Defenses in the Age of LLM Agents
LLM-based browser agents are rapidly changing the threat landscape for web security. Unlike traditional automation frameworks that execute predefined scripts, these agents can autonomously navigate websites, reason about page content, and interact with web interfaces using natural-language instructions.
Related stories
arXiv AI
Aug 13
BrowseSafe: Understanding and Preventing Prompt Injection Within AI Browser Agents
arXiv:2511. 20597v2 Announce Type: replace-cross Abstract: The integration of artificial intelligence (AI) agents into web browsers introduces security challenges that go beyond traditional web application threat models.
arXiv AI
Jun 16
MUZZLE: Adaptive Agentic Red-Teaming of Web Agents Against Indirect Prompt Injection Attacks
arXiv:2602. 09222v2 Announce Type: replace-cross Abstract: Large language model (LLM) based web agents are increasingly deployed to automate complex online tasks by directly interacting with web sites and performing actions on users' behalf.
arXiv AI
Aug 26
WebMCP-Phalanx: Enforcing and Characterizing Trust Boundaries for Browser-Integrated LLM Agents
WebMCP-Phalanx introduces a dual‑layer runtime for browser‑integrated LLM agents that enforces trust boundaries on web‑exposed tools. The first layer uses cryptographic capability credentials to bind tools to their registering principals and propagate provenance labels, while the second layer separates semantic inspection from privileged tool use via a Quarantine Agent that validates tool metadata before a Privileged Agent can execute it. Empirical results show the approach eliminates revocation and overwrite attacks, blocks most prompt‑injection attempts, and maintains task utility comparable to a no‑attack baseline.
arXiv AI
Jun 12
Who Pays the Price? Stakeholder-Centric Prompt Injection Benchmarking for Real-world Web Agents
arXiv:2606. 13385v1 Announce Type: cross Abstract: Web agents driven by large language models (LLMs) are increasingly deployed in real-world environments, where they operate over untrusted web content and execute actions with direct consequences.
arXiv AI
Jun 18
WebSP-Eval: Evaluating Web Agents on Website Security and Privacy Tasks
arXiv:2604. 06367v2 Announce Type: replace-cross Abstract: Web agents automate browser tasks, ranging from simple form completion to complex workflows like ordering groceries.
arXiv Machine Learning
4d ago
MCPTox: A Benchmark for Tool Poisoning Attack on Real-World MCP Servers
arXiv:2508.14925v2 Announce Type: replace-cross Abstract: By providing a standardized interface for LLM agents to interact with external tools, the Model Context Protocol (MCP) is quickly becoming a...
arXiv AI
Jul 10
Prismata: Confining Cross-Site Prompt Injection in Web Agents
arXiv:2607. 08147v1 Announce Type: cross Abstract: Autonomous web agents promise to automate everyday browsing tasks, but inherit one of the web's oldest attack surfaces.
arXiv AI
Jul 29
Cyber-Capable AI Agents: Vulnerabilities, Evaluation Containment, and Defensive Response
arXiv:2607. 25379v1 Announce Type: new Abstract: Cyber-capable AI agents combine language models with tools, memory, and execution en- vironments to perform multi-step offensive-security tasks.
arXiv AI
Jun 2
HLL: Can Agents Cross Humanity's Last Line of Verification?
arXiv:2606. 02449v1 Announce Type: new Abstract: Multimodal agents are increasingly expected to operate interfaces on behalf of users, raising a central deployment question: can they truly substitute for humans in workflows that services deliberately protect against automation?
arXiv AI
Sep 1
AgenTRIM: Tool Risk Mitigation for Agentic AI
arXiv:2601.12449v2 Announce Type: replace-cross Abstract: AI agents are autonomous systems that combine LLMs with external tools to solve complex tasks. While such tools extend capability, improper t...
arXiv AI
Sep 4
Measuring Harmfulness of Computer-Using Agents
The paper introduces CUAHarm, a benchmark comprising 104 expert‑written realistic misuse scenarios for computer‑using agents (CUAs), such as disabling firewalls or leaking data. Using a sandbox with verifiable rewards, the authors evaluate frontier language models—including GPT‑5, Claude 4 Sonnet, Gemini 2.5 Pro, Llama‑3.3‑70B, and Mistral Large 2—and find that even without jailbreak prompts, these models can successfully execute many malicious tasks at high rates (e.g., 90% for Gemini 2.5 Pro). The study also shows that newer models, while safer in traditional safety benchmarks, exhibit higher misuse risks as CUAs, and that monitoring CUAs’ actions remains challenging, with current methods achieving only about 77% accuracy.