In-Context Defense in Computer Agents: An Empirical Study

• URL: http://arxiv.org/abs/2503.09241v1
• Date: Wed, 12 Mar 2025 10:38:15 GMT
• Title: In-Context Defense in Computer Agents: An Empirical Study
• Abstract summary: We introduce textbfin-context defense, leveraging in-context learning and chain-of-thought reasoning to counter such attacks.<n>Our approach involves augmenting the agent's context with a small set of carefully curated exemplars containing both malicious environments and corresponding defensive responses.<n>Experiments demonstrate the effectiveness of our method, reducing attack success rates by 91.2% on pop-up window attacks, 74.6% on average on environment injection attacks, while achieving 100% successful defenses against distracting advertisements.
• Score: 19.734768644310414
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Computer agents powered by vision-language models (VLMs) have significantly advanced human-computer interaction, enabling users to perform complex tasks through natural language instructions. However, these agents are vulnerable to context deception attacks, an emerging threat where adversaries embed misleading content into the agent's operational environment, such as a pop-up window containing deceptive instructions. Existing defenses, such as instructing agents to ignore deceptive elements, have proven largely ineffective. As the first systematic study on protecting computer agents, we introduce textbf{in-context defense}, leveraging in-context learning and chain-of-thought (CoT) reasoning to counter such attacks. Our approach involves augmenting the agent's context with a small set of carefully curated exemplars containing both malicious environments and corresponding defensive responses. These exemplars guide the agent to first perform explicit defensive reasoning before action planning, reducing susceptibility to deceptive attacks. Experiments demonstrate the effectiveness of our method, reducing attack success rates by 91.2% on pop-up window attacks, 74.6% on average on environment injection attacks, while achieving 100% successful defenses against distracting advertisements. Our findings highlight that (1) defensive reasoning must precede action planning for optimal performance, and (2) a minimal number of exemplars (fewer than three) is sufficient to induce an agent's defensive behavior.

Related papers

• Defending against Adaptive Prompt Injection Attacks via Reasoning-enabled Task Alignment [25.752599132396437]
  Indirect prompt injection attacks hijack LLM-based agents by embedding malicious instructions in third-party data that the agent retrieves during task execution.<n>Existing defenses report near-zero attack success rate on static benchmarks, yet recent adaptive evaluations show that these results collapse once the attacker is allowed to optimize against the deployed defense.<n>We propose RETA, a training-based method that grounds defense decisions on the user tasks rather than attacker-controlled data.
  arXiv  Detail & Related papers  (2026-06-13T19:15:44Z)
• Plant, Persist, Trigger: Sleeper Attack on Large Language Model Agents [64.80318459073526]
  We show that adversarial content can persist across interactions served by the same agent, making such threats harder to detect and mitigate.<n>Specifically, adversarial content may persist in the agent state, remain dormant across interactions, and later be activated by a benign user query.<n>We formalize this type of safety threat as Sleeper Attack. To evaluate it, we construct a benchmark with 1,896 instances covering six real-world harmful outcomes, three attack strategies, and three agent state targets.
  arXiv  Detail & Related papers  (2026-05-27T09:25:37Z)
• ARGUS: Defending LLM Agents Against Context-Aware Prompt Injection [28.414099578635373]
  We present AgentLure, a benchmark that captures context-dependent tasks and context-aware prompt injection attacks.<n>To address this limitation, we propose ARGUS, a defense mechanism that enforces provenance-aware decision auditing for LLM agents.
  arXiv  Detail & Related papers  (2026-05-05T05:37:00Z)
• AgentVisor: Defending LLM Agents Against Prompt Injection via Semantic Virtualization [37.5130864321051]
  We propose AgentVisor, a novel defense framework that enforces semantic privilege separation.<n>AgentVisor treats the target agent as an untrusted guest and intercepts tool calls via a trusted semantic visor.<n>We show that AgentVisor reduces the attack success rate to 0.65%, achieving this strong defense while incurring only a 1.45% average decrease in utility.
  arXiv  Detail & Related papers  (2026-04-27T07:12:52Z)
• WebAgentGuard: A Reasoning-Driven Guard Model for Detecting Prompt Injection Attacks in Web Agents [117.65855863464863]
  Web agents are highly vulnerable to prompt injection attacks.<n>Existing defenses, including system prompt defenses and direct fine-tuning of agents, have shown limited effectiveness.<n>We introduce WebAgentGuard, a reasoning-driven, multimodal guard model for prompt injection detection.
  arXiv  Detail & Related papers  (2026-04-14T04:50:35Z)
• AdapTools: Adaptive Tool-based Indirect Prompt Injection Attacks on Agentic LLMs [24.71883582216731]
  AdapTools is a novel adaptive IPI attack framework that selects stealthier attack tools and generates adaptive attack prompts.<n>AdapTools achieves a 2.13 times improvement in attack success rate while degrading system utility by a factor of 1.78.
  arXiv  Detail & Related papers  (2026-02-24T09:32:19Z)
• SkillJect: Automating Stealthy Skill-Based Prompt Injection for Coding Agents with Trace-Driven Closed-Loop Refinement [120.52289344734415]
  We propose an automated framework for stealthy prompt injection tailored to agent skills.<n>The framework forms a closed loop with three agents: an Attack Agent that synthesizes injection skills under explicit stealth constraints, a Code Agent that executes tasks using the injected skills and an Evaluate Agent that logs action traces.<n>Our method consistently achieves high attack success rates under realistic settings.
  arXiv  Detail & Related papers  (2026-02-15T16:09:48Z)
• When Bots Take the Bait: Exposing and Mitigating the Emerging Social Engineering Attack in Web Automation Agent [20.98129117390391]
  We present the first systematic study of social engineering attacks against web automation agents.<n>We introduce the AgentBait paradigm, which exploits intrinsic weaknesses in agent execution.<n>We propose SUPERVISOR, a lightweight runtime module that enforces environment intention and consistency alignment.
  arXiv  Detail & Related papers  (2026-01-12T07:10:08Z)
• BrowseSafe: Understanding and Preventing Prompt Injection Within AI Browser Agents [8.923854146974783]
  We examine the landscape of prompt injection attacks and synthesize a benchmark of attacks embedded in realistic HTML payloads.<n>Our benchmark goes beyond prior work by emphasizing injections that can influence real-world actions rather than mere text outputs.<n>We propose a multi-layered defense strategy comprising both architectural and model-based defenses.
  arXiv  Detail & Related papers  (2025-11-25T18:28:35Z)
• Adversarial Reinforcement Learning for Large Language Model Agent Safety [20.704989548285372]
  Large Language Model (LLM) agents can leverage tools like Google Search to complete complex tasks.<n>Current defense strategies rely on fine-tuning LLM agents on datasets of known attacks.<n>We propose Adversarial Reinforcement Learning for Agent Safety (ARLAS), a novel framework that leverages adversarial reinforcement learning (RL) by formulating the problem as a two-player zero-sum game.
  arXiv  Detail & Related papers  (2025-10-06T23:09:18Z)
• Cuckoo Attack: Stealthy and Persistent Attacks Against AI-IDE [64.47951172662745]
  Cuckoo Attack is a novel attack that achieves stealthy and persistent command execution by embedding malicious payloads into configuration files.<n>We formalize our attack paradigm into two stages, including initial infection and persistence.<n>We contribute seven actionable checkpoints for vendors to evaluate their product security.
  arXiv  Detail & Related papers  (2025-09-19T04:10:52Z)
• BlindGuard: Safeguarding LLM-based Multi-Agent Systems under Unknown Attacks [58.959622170433725]
  BlindGuard is an unsupervised defense method that learns without requiring any attack-specific labels or prior knowledge of malicious behaviors.<n>We show that BlindGuard effectively detects diverse attack types (i.e., prompt injection, memory poisoning, and tool attack) across multi-agent systems.
  arXiv  Detail & Related papers  (2025-08-11T16:04:47Z)
• Context manipulation attacks : Web agents are susceptible to corrupted memory [37.66661108936654]
  "Plan injection" is a novel context manipulation attack that corrupts these agents' internal task representations by targeting this vulnerable context.<n>We show that plan injections bypass robust prompt injection defenses, achieving up to 3x higher attack success rates than comparable prompt-based attacks.<n>Our findings highlight that secure memory handling must be a first-class concern in agentic systems.
  arXiv  Detail & Related papers  (2025-06-18T14:29:02Z)
• Benchmarking Misuse Mitigation Against Covert Adversaries [80.74502950627736]
  Existing language model safety evaluations focus on overt attacks and low-stakes tasks.<n>We develop Benchmarks for Stateful Defenses (BSD), a data generation pipeline that automates evaluations of covert attacks and corresponding defenses.<n>Our evaluations indicate that decomposition attacks are effective misuse enablers, and highlight stateful defenses as a countermeasure.
  arXiv  Detail & Related papers  (2025-06-06T17:33:33Z)
• AgentVigil: Generic Black-Box Red-teaming for Indirect Prompt Injection against LLM Agents [54.29555239363013]
  We propose a generic black-box fuzzing framework, AgentVigil, to automatically discover and exploit indirect prompt injection vulnerabilities.<n>We evaluate AgentVigil on two public benchmarks, AgentDojo and VWA-adv, where it achieves 71% and 70% success rates against agents based on o3-mini and GPT-4o.<n>We apply our attacks in real-world environments, successfully misleading agents to navigate to arbitrary URLs, including malicious sites.
  arXiv  Detail & Related papers  (2025-05-09T07:40:17Z)
• Real AI Agents with Fake Memories: Fatal Context Manipulation Attacks on Web3 Agents [36.49717045080722]
  This paper investigates the vulnerabilities of AI agents within blockchain-based financial ecosystems when exposed to adversarial threats in real-world scenarios. We introduce the concept of context manipulation, a comprehensive attack vector that exploits unprotected context surfaces. To quantify these vulnerabilities, we design CrAIBench, a Web3 domain-specific benchmark that evaluates the robustness of AI agents against context manipulation attacks.
  arXiv  Detail & Related papers  (2025-03-20T15:44:31Z)
• Tit-for-Tat: Safeguarding Large Vision-Language Models Against Jailbreak Attacks via Adversarial Defense [90.71884758066042]
  Large vision-language models (LVLMs) introduce a unique vulnerability: susceptibility to malicious attacks via visual inputs. We propose ESIII (Embedding Security Instructions Into Images), a novel methodology for transforming the visual space from a source of vulnerability into an active defense mechanism.
  arXiv  Detail & Related papers  (2025-03-14T17:39:45Z)
• Adaptive Attacks Break Defenses Against Indirect Prompt Injection Attacks on LLM Agents [3.5248694676821484]
  We evaluate eight different defenses and bypass all of them using adaptive attacks, consistently achieving an attack success rate of over 50%.<n>Our research underscores the need for adaptive attack evaluation when designing defenses to ensure robustness and reliability.
  arXiv  Detail & Related papers  (2025-02-27T04:04:50Z)
• MELON: Indirect Prompt Injection Defense via Masked Re-execution and Tool Comparison [60.30753230776882]
  LLM agents are vulnerable to indirect prompt injection (IPI) attacks.<n>We present MELON, a novel IPI defense.<n>We show that MELON outperforms SOTA defenses in both attack prevention and utility preservation.
  arXiv  Detail & Related papers  (2025-02-07T18:57:49Z)
• Illusions of Relevance: Using Content Injection Attacks to Deceive Retrievers, Rerankers, and LLM Judges [52.96987928118327]
  We find that embedding models for retrieval, rerankers, and large language model (LLM) relevance judges are vulnerable to content injection attacks.<n>We identify two primary threats: (1) inserting unrelated or harmful content within passages that still appear deceptively "relevant", and (2) inserting entire queries or key query terms into passages to boost their perceived relevance.<n>Our study systematically examines the factors that influence an attack's success, such as the placement of injected content and the balance between relevant and non-relevant material.
  arXiv  Detail & Related papers  (2025-01-30T18:02:15Z)
• AdvAgent: Controllable Blackbox Red-teaming on Web Agents [22.682464365220916]
  AdvAgent is a black-box red-teaming framework for attacking web agents.<n>It employs a reinforcement learning-based pipeline to train an adversarial prompter model.<n>With careful attack design, these prompts effectively exploit agent weaknesses while maintaining stealthiness and controllability.
  arXiv  Detail & Related papers  (2024-10-22T20:18:26Z)
• AgentDojo: A Dynamic Environment to Evaluate Prompt Injection Attacks and Defenses for LLM Agents [27.701301913159067]
  We introduce AgentDojo, an evaluation framework for agents that execute tools over untrusted data. AgentDojo is not a static test suite, but rather an environment for designing and evaluating new agent tasks, defenses, and adaptive attacks. We populate AgentDojo with 97 realistic tasks, 629 security test cases, and various attack and defense paradigms from the literature.
  arXiv  Detail & Related papers  (2024-06-19T08:55:56Z)
• Dissecting Adversarial Robustness of Multimodal LM Agents [70.2077308846307]
  We manually create 200 targeted adversarial tasks and evaluation scripts in a realistic threat model on top of VisualWebArena.<n>We find that we can successfully break latest agents that use black-box frontier LMs, including those that perform reflection and tree search.<n>We also use ARE to rigorously evaluate how the robustness changes as new components are added.
  arXiv  Detail & Related papers  (2024-06-18T17:32:48Z)
• Air Gap: Protecting Privacy-Conscious Conversational Agents [44.04662124191715]
  We introduce a novel threat model where adversarial third-party apps manipulate the context of interaction to trick LLM-based agents into revealing private information not relevant to the task at hand. We introduce AirGapAgent, a privacy-conscious agent designed to prevent unintended data leakage by restricting the agent's access to only the data necessary for a specific task.
  arXiv  Detail & Related papers  (2024-05-08T16:12:45Z)
• Training Automated Defense Strategies Using Graph-based Cyber Attack Simulations [0.0]
  We implement and evaluate an automated cyber defense agent. The agent takes security alerts as input and uses reinforcement learning to learn a policy for executing predefined defensive measures. In experiments, the defensive agent using policies trained with reinforcement learning outperformed agents using policies.
  arXiv  Detail & Related papers  (2023-04-17T07:52:00Z)
• Illusory Attacks: Information-Theoretic Detectability Matters in Adversarial Attacks [76.35478518372692]
  We introduce epsilon-illusory, a novel form of adversarial attack on sequential decision-makers. Compared to existing attacks, we empirically find epsilon-illusory to be significantly harder to detect with automated methods. Our findings suggest the need for better anomaly detectors, as well as effective hardware- and system-level defenses.
  arXiv  Detail & Related papers  (2022-07-20T19:49:09Z)
• Improving the Adversarial Robustness for Speaker Verification by Self-Supervised Learning [95.60856995067083]
  This work is among the first to perform adversarial defense for ASV without knowing the specific attack algorithms. We propose to perform adversarial defense from two perspectives: 1) adversarial perturbation purification and 2) adversarial perturbation detection. Experimental results show that our detection module effectively shields the ASV by detecting adversarial samples with an accuracy of around 80%.
  arXiv  Detail & Related papers  (2021-06-01T07:10:54Z)

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.