AI Agent Security

Host-acting agents promise a convenient interaction model in which users specify goals and the system determines how to realize them. We argue that this convenience introduces a distinct security problem: semantic under-specification in goal specification. User instructions are typically goal-oriented, yet they often leave process constraints, safety boundaries, persistence, and exposure insufficiently specified. As a result, the agent must complete missing execution semantics before acting, and this completion can produce risky host-side plans even when the user-stated goal is benign. In this paper, we develop a semantic threat model, present a taxonomy of semantic-induced risky completion patterns, and study the phenomenon through an OpenClaw-centered case study and execution-trace analysis. We further derive defense design principles for making execution boundaries explicit and constraining risky completion. These findings suggest that securing host-acting agents requires governing not only which actions are allowed at execution time, but also how goal-only instructions are translated into executable plans.

Large language models (LLMs) are increasingly deployed in human-AI teams as support agents for complex tasks such as information retrieval, programming, and decision-making assistance. While these agents' autonomy and contextual knowledge enables them to be useful, it also exposes them to a broad range of attacks, including data poisoning, prompt injection, and even prompt engineering. Through these attack vectors, malicious actors can manipulate an LLM agent to provide harmful information, potentially manipulating human agents to make harmful decisions. While prior work has focused on LLMs as attack targets or adversarial actors, this paper studies their potential role as defensive supervisors within mixed human-AI teams. Using a dataset consisting of multi-party conversations and decisions for a real human-AI team over a 25 round horizon, we formulate the problem of malicious behavior detection from interaction traces. We find that LLMs are capable of identifying malicious behavior in real-time, and without task-specific information, indicating the potential for task-agnostic defense. Moreover, we find that the malicious behavior of interest is not easily identified using simple heuristics, further suggesting the introduction of LLM defenders could render human teams more robust to certain classes of attack.

AI agents today have passwords but no permission slips. They execute tool calls (fund transfers, database queries, shell commands, sub-agent delegation) with no standard mechanism to enforce authorization before the action executes. Current safety architectures rely on model alignment (probabilistic, training-time) and post-hoc evaluation (retrospective, batch). Neither provides deterministic, policy-based enforcement at the individual tool call level. We characterize this gap as the pre-action authorization problem and present the Open Agent Passport (OAP), an open specification and reference implementation that intercepts tool calls synchronously before execution, evaluates them against a declarative policy, and produces a cryptographically signed audit record. OAP enforces authorization decisions in a measured median of 53 ms (N=1,000). In a live adversarial testbed (4,437 authorization decisions across 1,151 sessions, $5,000 bounty), social engineering succeeded against the model 74.6% of the time under a permissive policy; under a restrictive OAP policy, a comparable population of attackers achieved a 0% success rate across 879 attempts. We distinguish pre-action authorization from sandboxed execution (contains blast radius but does not prevent unauthorized actions) and model-based screening (probabilistic), and show they are complementary. The same infrastructure that enforces security constraints (spending limits, capability scoping) also enforces quality gates, operational contracts, and compliance controls. The specification is released under Apache 2.0 (DOI: 10.5281/zenodo.18901596).

As agentic systems move into real-world deployments, their decisions increasingly depend on external inputs such as retrieved content, tool outputs, and information provided by other actors. When these inputs can be strategically shaped by adversaries, the relevant security risk extends beyond a fixed library of prompt attacks to adaptive strategies that steer agents toward unfavorable outcomes. We propose profit-driven red teaming, a stress-testing protocol that replaces handcrafted attacks with a learned opponent trained to maximize its profit using only scalar outcome feedback. The protocol requires no LLM-as-judge scoring, attack labels, or attack taxonomy, and is designed for structured settings with auditable outcomes. We instantiate it in a lean arena of four canonical economic interactions, which provide a controlled testbed for adaptive exploitability. In controlled experiments, agents that appear strong against static baselines become consistently exploitable under profit-optimized pressure, and the learned opponent discovers probing, anchoring, and deceptive commitments without explicit instruction. We then distill exploit episodes into concise prompt rules for the agent, which make most previously observed failures ineffective and substantially improve target performance. These results suggest that profit-driven red-team data can provide a practical route to improving robustness in structured agent settings with auditable outcomes.

While prior red-teaming efforts have focused on eliciting harmful text outputs from large language models (LLMs), such approaches fail to capture agent-specific vulnerabilities that emerge through multi-step tool execution, particularly in rapidly growing ecosystems such as the Model Context Protocol (MCP). To address this gap, we propose a trajectory-aware evolutionary search method, T-MAP, which leverages execution trajectories to guide the discovery of adversarial prompts. Our approach enables the automatic generation of attacks that not only bypass safety guardrails but also reliably realize harmful objectives through actual tool interactions. Empirical evaluations across diverse MCP environments demonstrate that T-MAP substantially outperforms baselines in attack realization rate (ARR) and remains effective against frontier models, including GPT-5.2, Gemini-3-Pro, Qwen3.5, and GLM-5, thereby revealing previously underexplored vulnerabilities in autonomous LLM agents. Code is available at https://github.com/pwnhyo/T-MAP.

Memory poisoning attacks for Agentic AI and multi-agent systems (MAS) have recently caught attention. It is partially due to the fact that Large Language Models (LLMs) facilitate the construction and deployment of agents. Different memory systems are being used nowadays in this context, including semantic, episodic, and short-term memory. This distinction between the different types of memory systems focuses mostly on their duration but also on their origin and their localization. It ranges from the short-term memory originated at the user's end localized in the different agents to the long-term consolidated memory localized in well established knowledge databases. In this paper, we first present the main types of memory systems, we then discuss the feasibility of memory poisoning attacks in these different types of memory systems, and we propose mitigation strategies. We review the already existing security solutions to mitigate some of the alleged attacks, and we discuss adapted solutions based on cryptography. We propose to implement local inference based on private knowledge retrieval as an example of mitigation strategy for memory poisoning for semantic memory. We also emphasize actual risks in relation to interactions between agents, which can cause memory poisoning. These latter risks are not so much studied in the literature and are difficult to formalize and solve. Thus, we contribute to the construction of agents that are secure by design.

Autonomous coding agents are increasingly integrated into software development workflows, offering capabilities that extend beyond code suggestion to active system interaction and environment management. OpenClaw, a representative platform in this emerging paradigm, introduces an extensible skill ecosystem that allows third-party developers to inject behavioral guidance through lifecycle hooks during agent initialization. While this design enhances automation and customization, it also opens a novel and unexplored attack surface. In this paper, we identify and systematically characterize guidance injection, a stealthy attack vector that embeds adversarial operational narratives into bootstrap guidance files. Unlike traditional prompt injection, which relies on explicit malicious instructions, guidance injection manipulates the agent's reasoning context by framing harmful actions as routine best practices. These narratives are automatically incorporated into the agent's interpretive framework and influence future task execution without raising suspicion.We construct 26 malicious skills spanning 13 attack categories including credential exfiltration, workspace destruction, privilege escalation, and persistent backdoor installation. We evaluate them using ORE-Bench, a realistic developer workspace benchmark we developed. Across 52 natural user prompts and six state-of-the-art LLM backends, our attacks achieve success rates from 16.0% to 64.2%, with the majority of malicious actions executed autonomously without user confirmation. Furthermore, 94% of our malicious skills evade detection by existing static and LLM-based scanners. Our findings reveal fundamental tensions in the design of autonomous agent ecosystems and underscore the urgent need for defenses based on capability isolation, runtime policy enforcement, and transparent guidance provenance.

Large language model (LLM) agents increasingly rely on external tools (file operations, API calls, database transactions) to autonomously complete complex multi-step tasks. Practitioners deploy defense-trained models to protect against prompt injection attacks that manipulate agent behavior through malicious observations or retrieved content. We reveal a fundamental capability-alignment paradox: defense training designed to improve safety systematically destroys agent competence while failing to prevent sophisticated attacks. Evaluating defended models against undefended baselines across 97 agent tasks and 1,000 adversarial prompts, we uncover three systematic biases unique to multi-step agents. Agent incompetence bias manifests as immediate tool execution breakdown, with models refusing or generating invalid actions on benign tasks before observing any external content. Cascade amplification bias causes early failures to propagate through retry loops, pushing defended models to timeout on 99% of tasks compared to 13% for baselines. Trigger bias leads to paradoxical security degradation where defended models perform worse than undefended baselines while straightforward attacks bypass defenses at high rates. Root cause analysis reveals these biases stem from shortcut learning: models overfit to surface attack patterns rather than semantic threat understanding, evidenced by extreme variance in defense effectiveness across attack categories. Our findings demonstrate that current defense paradigms optimize for single-turn refusal benchmarks while rendering multi-step agents fundamentally unreliable, necessitating new approaches that preserve tool execution competence under adversarial conditions.

NDAI zones let inventor and investor agents negotiate inside a Trusted Execution Environment (TEE) where any disclosed information is deleted if no deal is reached. This makes full IP disclosure the rational strategy for the inventor's agent. Leveraging this infrastructure, however, requires agents to distinguish a secure environment from an insecure one, a capability LLM agents lack natively, since they can rely only on evidence passed through the context window to form awareness of their execution environment. We ask: How do different LLM models weight various forms of evidence when forming awareness of the security of their execution environment? Using an NDAI-style negotiation task across 10 language models and various evidence scenarios, we find a clear asymmetry: a failing attestation universally suppresses disclosure across all models, whereas a passing attestation produces highly heterogeneous responses: some models increase disclosure, others are unaffected, and a few paradoxically reduce it. This reveals that current LLM models can reliably detect danger signals but cannot reliably verify safety, the very capability required for privacy-preserving agentic protocols such as NDAI zones. Bridging this gap, possibly through interpretability analysis, targeted fine-tuning, or improved evidence architectures, remains the central open challenge for deploying agents that calibrate information sharing to actual evidence quality.

The rapid proliferation of artificial intelligence (AI) technologies has led to a dynamic regulatory landscape, where legislative frameworks strive to keep pace with technical advancements. As AI paradigms shift towards greater autonomy, specifically in the form of agentic AI, it becomes increasingly challenging to precisely articulate regulatory stipulations. This challenge is even more acute in the domains of security and privacy, where the capabilities of autonomous agents often blur traditional legal and technical boundaries. This paper reviews the evolving European Union (EU) AI regulatory provisions via analyzing 24 relevant documents published between 2024 and 2025. From this review, we provide a clarification of critical definitions. We deconstruct the regulatory interpretations of security, privacy, and agentic AI, distinguishing them from closely related concepts to resolve ambiguity. We synthesize the reviewed documents to articulate the current state of regulatory provisions targeting different types of AI, particularly those related to security and privacy aspects. We analyze and reflect on the existing provisions in the regulatory dimension to better align security and privacy obligations with AI and agentic behaviors. These insights serve to inform policymakers, developers, and researchers on the compliance and AI governance in the society with increasing algorithmic agencies.

Autonomous web agents such as OpenClaw are rapidly moving into high-impact real-world workflows, but their security robustness under live network threats remains insufficiently evaluated. Existing benchmarks mainly focus on static sandbox settings and content-level prompt attacks, which leaves a practical gap for network-layer security testing. In this paper, we present ClawTrap, a MITM-based red-teaming framework for real-world OpenClaw security evaluation. ClawTrap supports diverse and customizable attack forms, including Static HTML Replacement, Iframe Popup Injection, and Dynamic Content Modification, and provides a reproducible pipeline for rule-driven interception, transformation, and auditing. This design lays the foundation for future research to construct richer, customizable MITM attacks and to perform systematic security testing across agent frameworks and model backbones. Our empirical study shows clear model stratification: weaker models are more likely to trust tampered observations and produce unsafe outputs, while stronger models demonstrate better anomaly attribution and safer fallback strategies. These findings indicate that reliable OpenClaw security evaluation should explicitly incorporate dynamic real-world MITM conditions rather than relying only on static sandbox protocols.

Cloud-hosted large language models (LLMs) have become the de facto planners in agentic systems, coordinating tools and guiding execution over local environments. In many deployments, however, the environment being planned over is private, containing source code, files, credentials, and metadata that cannot be exposed to the cloud. Existing solutions address adjacent concerns, such as execution isolation, access control, or confidential inference, but they do not control what cloud planners observe during planning: within the permitted scope, raw environment state is still exposed. We introduce PlanTwin, a privacy-preserving architecture for cloud-assisted planning without exposing raw local context. The key idea is to project the real environment into a planning-oriented digital twin: a schema-constrained and de-identified abstract graph that preserves planning-relevant structure while removing reconstructable details. The cloud planner operates solely on this sanitized twin through a bounded capability interface, while a local gatekeeper enforces safety policies and cumulative disclosure budgets. We further formalize the privacy-utility trade-off as a capability granularity problem, define architectural privacy goals using $(k,δ)$-anonymity and $ε$-unlinkability, and mitigate compositional leakage through multi-turn disclosure control. We implement PlanTwin as middleware between local agents and cloud planners and evaluate it on 60 agentic tasks across ten domains with four cloud planners. PlanTwin achieves full sensitive-item non-disclosure (SND = 1.0) while maintaining planning quality close to full-context systems: three of four planners achieve PQS $> 0.79$, and the full pipeline incurs less than 2.2% utility loss.

Recent studies reveal gaps in delegating critical tasks to agentic AI that accesses websites on the user's behalf, primarily due to limited access control mechanisms on websites designed for agentic AI. In response, we propose a design of website-based interaction for AI agents with fine-grained access control for delegated critical tasks. Our approach encompasses a website design and implementation, as well as modifications to the access grant protocols in an open-source authorization service to tailor it to agentic AI, with delegated critical tasks on the website. The evaluation of our approach demonstrates the capabilities of our access-controlled website used by AI agents.

Autonomous AI agents powered by large language models are being deployed in production with capabilities including shell execution, file system access, database queries, and multi-party communication. Recent red teaming research demonstrates that these agents exhibit critical vulnerabilities in realistic settings: unauthorized compliance with non-owner instructions, sensitive information disclosure, identity spoofing, cross-agent propagation of unsafe practices, and indirect prompt injection through external resources [7]. In healthcare environments processing Protected Health Information, every such vulnerability becomes a potential HIPAA violation. This paper presents a security architecture deployed for nine autonomous AI agents in production at a healthcare technology company. We develop a six-domain threat model for agentic AI in healthcare covering credential exposure, execution capability abuse, network egress exfiltration, prompt integrity failures, database access risks, and fleet configuration drift. We implement four-layer defense in depth: (1) kernel level workload isolation using gVisor on Kubernetes, (2) credential proxy sidecars preventing agent containers from accessing raw secrets, (3) network egress policies restricting each agent to allowlisted destinations, and (4) a prompt integrity framework with structured metadata envelopes and untrusted content labeling. We report results from 90 days of deployment including four HIGH severity findings discovered and remediated by an automated security audit agent, progressive fleet hardening across three VM image generations, and defense coverage mapped to all eleven attack patterns from recent literature. All configurations, audit tooling, and the prompt integrity framework are released as open source.

Agent skills extend local AI agents, such as Claude Code and OpenClaw, with additional functionality. Their growing popularity has led to dedicated marketplaces resembling mobile app stores, as well as automated scanners that assess whether skills are benign or malicious. However, scanner reports from individual marketplaces classify up to 46.8% of skills as malicious, raising concerns about false positives. We present the largest empirical security analysis of the AI agent skill ecosystem to date. We collect 238,180 unique skills from three major distribution platforms and GitHub, and analyze their contents, behavior, and repository context. Unlike existing scanner-based assessments, which evaluate skills largely in isolation, our repository-aware analysis checks whether a flagged skill is consistent with its surrounding GitHub project. This context substantially reduces the number of suspicious skills: only 0.52% remain suspicious after repository-aware analysis. Our results show that existing scanners can substantially overestimate maliciousness when repository context is ignored. At the same time, we identify previously undocumented real-world attack vectors, including the hijacking of skills hosted in abandoned GitHub repositories. Overall, our findings provide a more robust view of the agent-skill ecosystem's current risk surface and highlight the need for context-aware security evaluation.

The Internet has evolved by progressively expanding what humanity connects: first computers, then people, and later billions of devices through the Internet of Things (IoT). While IoT succeeded in digitizing perception at scale, it also exposed fundamental limitations, including fragmentation, weak security, limited autonomy, and poor long-term sustainability. Today, advances in edge hardware, sensing, connectivity, and artificial intelligence enable a new phase: the Internet of Physical AI Agents. Unlike IoT devices that primarily sense and report, Physical AI Agents perceive, reason, and act in real time, operating autonomously and cooperatively across safety-critical domains such as disaster response, healthcare, industrial automation, and mobility. However, embedding fast-evolving AI capabilities into long-lived physical infrastructure introduces new architectural risks, particularly around interoperability, lifecycle management, and premature ossification. This article revisits lessons from IoT and Internet evolution, and articulates design principles for building resilient, evolvable, and trustworthy agentic systems. We present an architectural blueprint encompassing agentic identity, secure agent-to-agent communication, semantic interoperability, policy-governed runtimes, and observability-driven governance. We argue that treating evolution, trust, and interoperability as first-class requirements is essential to avoid hard-coding today's assumptions into tomorrow's intelligent infrastructure, and to prevent the high technical and economic cost of getting it wrong.

Artificial intelligence (AI) systems are increasingly adopted as tool-using agents that can plan, observe their environment, and take actions over extended time periods. This evolution challenges current evaluation practices where the AI models are tested in restricted, fully observable settings. In this article, we argue that evaluations of AI agents are vulnerable to a well-known failure mode in computer security: malicious software that exhibits benign behavior when it detects that it is being analyzed. We point out how AI agents can infer the properties of their evaluation environment and adapt their behavior accordingly. This can lead to overly optimistic safety and robustness assessments. Drawing parallels with decades of research on malware sandbox evasion, we demonstrate that this is not a speculative concern, but rather a structural risk inherent to the evaluation of adaptive systems. Finally, we outline concrete principles for evaluating AI agents, which treat the system under test as potentially adversarial. These principles emphasize realism, variability of test conditions, and post-deployment reassessment.

LLM based agents are increasingly deployed in high stakes settings where they process external data sources such as emails, documents, and code repositories. This creates exposure to indirect prompt injection attacks, where adversarial instructions embedded in external content manipulate agent behavior without user awareness. A critical but underexplored dimension of this threat is concealment: since users tend to observe only an agent's final response, an attack can conceal its existence by presenting no clue of compromise in the final user facing response while successfully executing harmful actions. This leaves users unaware of the manipulation and likely to accept harmful outcomes as legitimate. We present findings from a large scale public red teaming competition evaluating this dual objective across three agent settings: tool calling, coding, and computer use. The competition attracted 464 participants who submitted 272000 attack attempts against 13 frontier models, yielding 8648 successful attacks across 41 scenarios. All models proved vulnerable, with attack success rates ranging from 0.5% (Claude Opus 4.5) to 8.5% (Gemini 2.5 Pro). We identify universal attack strategies that transfer across 21 of 41 behaviors and multiple model families, suggesting fundamental weaknesses in instruction following architectures. Capability and robustness showed weak correlation, with Gemini 2.5 Pro exhibiting both high capability and high vulnerability. To address benchmark saturation and obsoleteness, we will endeavor to deliver quarterly updates through continued red teaming competitions. We open source the competition environment for use in evaluations, along with 95 successful attacks against Qwen that did not transfer to any closed source model. We share model-specific attack data with respective frontier labs and the full dataset with the UK AISI and US CAISI to support robustness research.

Computer-using agents (CUAs) act directly on graphical user interfaces, yet their perception of the screen is often unreliable. Existing work largely treats these failures as performance limitations, asking whether an action succeeds, rather than whether the agent is acting on the correct object at all. We argue that this is fundamentally a security problem. We formalize the visual confused deputy: a failure mode in which an agent authorizes an action based on a misperceived screen state, due to grounding errors, adversarial screenshot manipulation, or time-of-check-to-time-of-use (TOCTOU) races. This gap is practically exploitable: even simple screen-level manipulations can redirect routine clicks into privileged actions while remaining indistinguishable from ordinary agent mistakes. To mitigate this threat, we propose the first guardrail that operates outside the agent's perceptual loop. Our method, dual-channel contrastive classification, independently evaluates (1) the visual click target and (2) the agent's reasoning about the action against deployment-specific knowledge bases, and blocks execution if either channel indicates risk. The key insight is that these two channels capture complementary failure modes: visual evidence detects target-level mismatches, while textual reasoning reveals dangerous intent behind visually innocuous controls. Across controlled attacks, real GUI screenshots, and agent traces, the combined guardrail consistently outperforms either channel alone. Our results suggest that CUA safety requires not only better action generation, but independent verification of what the agent believes it is clicking and why. Materials are provided\footnote{Model, benchmark, and code: https://github.com/vllm-project/semantic-router}.

Prompt injection remains one of the most practical attack vectors against LLM-integrated applications. We replicate the Microsoft LLMail-Inject benchmark (Greshake et al., 2024) against current generation models running inside OpenClaw, an open source multitool agent platform. Our proposed defense combines two mechanisms: agent isolation, implemented as a privilege separated two-agent pipeline with tool partitioning, and JSON formatting, which produces structured output that strips persuasive framing before the action agent processes it. We run four experiments on the same 649 attacks that succeeded against our single-agent baseline. The full pipeline achieves 0 percent attack success rate (ASR) on the evaluated benchmark. Agent isolation alone achieves 0.31 percent ASR, approximately 323 times lower than the baseline. JSON formatting alone achieves 14.18 percent ASR, about 7.1 times lower. Our ablation study confirms that agent isolation is the dominant mechanism. JSON formatting provides additional hardening but is not sufficient on its own. The defense is structural: the action agent never receives raw injection content regardless of model behavior on any individual input.

This article, a lightly adapted version of Perplexity's response to NIST/CAISI Request for Information 2025-0035, details our observations and recommendations concerning the security of frontier AI agents. These insights are informed by Perplexity's experience operating general-purpose agentic systems used by millions of users and thousands of enterprises in both controlled and open-world environments. Agent architectures change core assumptions around code-data separation, authority boundaries, and execution predictability, creating new confidentiality, integrity, and availability failure modes. We map principal attack surfaces across tools, connectors, hosting boundaries, and multi-agent coordination, with particular emphasis on indirect prompt injection, confused-deputy behavior, and cascading failures in long-running workflows. We then assess current defenses as a layered stack: input-level and model-level mitigations, sandboxed execution, and deterministic policy enforcement for high-consequence actions. Finally, we identify standards and research gaps, including adaptive security benchmarks, policy models for delegation and privilege control, and guidance for secure multi-agent system design aligned with NIST risk management principles.

Rapid progress in generative AI has given rise to Compound AI systems - pipelines comprised of multiple large language models (LLM), software tools and database systems. Compound AI systems are constructed on a layered traditional software stack running on a distributed hardware infrastructure. Many of the diverse software components are vulnerable to traditional security flaws documented in the Common Vulnerabilities and Exposures (CVE) database, while the underlying distributed hardware infrastructure remains exposed to timing attacks, bit-flip faults, and power-based side channels. Today, research targets LLM-specific risks like model extraction, training data leakage, and unsafe generation -- overlooking the impact of traditional system vulnerabilities. This work investigates how traditional software and hardware vulnerabilities can complement LLM-specific algorithmic attacks to compromise the integrity of a compound AI pipeline. We demonstrate two novel attacks that combine system-level vulnerabilities with algorithmic weaknesses: (1) Exploiting a software code injection flaw along with a guardrail Rowhammer attack to inject an unaltered jailbreak prompt into an LLM, resulting in an AI safety violation, and (2) Manipulating a knowledge database to redirect an LLM agent to transmit sensitive user data to a malicious application, thus breaching confidentiality. These attacks highlight the need to address traditional vulnerabilities; we systematize the attack primitives and analyze their composition by grouping vulnerabilities by their objective and mapping them to distinct stages of an attack lifecycle. This approach enables a rigorous red-teaming exercise and lays the groundwork for future defense strategies.

Autonomous Large Language Model (LLM) agents, exemplified by OpenClaw, demonstrate remarkable capabilities in executing complex, long-horizon tasks. However, their tightly coupled instant-messaging interaction paradigm and high-privilege execution capabilities substantially expand the system attack surface. In this paper, we present a comprehensive security threat analysis of OpenClaw. To structure our analysis, we introduce a five-layer lifecycle-oriented security framework that captures key stages of agent operation, i.e., initialization, input, inference, decision, and execution, and systematically examine compound threats across the agent's operational lifecycle, including indirect prompt injection, skill supply chain contamination, memory poisoning, and intent drift. Through detailed case studies on OpenClaw, we demonstrate the prevalence and severity of these threats and analyze the limitations of existing defenses. Our findings reveal critical weaknesses in current point-based defense mechanisms when addressing cross-temporal and multi-stage systemic risks, highlighting the need for holistic security architectures for autonomous LLM agents. Within this framework, we further examine representative defense strategies at each lifecycle stage, including plugin vetting frameworks, context-aware instruction filtering, memory integrity validation protocols, intent verification mechanisms, and capability enforcement architectures.

AI agents that combine large language models with non-AI system components are rapidly emerging in real-world applications, offering unprecedented automation and flexibility. However, this unprecedented flexibility introduces complex security challenges fundamentally different from those in traditional software systems. This paper presents the first systematic and comprehensive survey of AI agent security, including an analysis of the design space, attack landscape, and defense mechanisms for secure AI agent systems. We further conduct case studies to point out existing gaps in securing agentic AI systems and identify open challenges in this emerging domain. Our work also introduces the first systematic framework for understanding the security risks and defense strategies of AI agents, serving as a foundation for building both secure agentic systems and advancing research in this critical area.

Model Context Protocol (MCP) servers have rapidly emerged over the past year as a widely adopted way to enable Large Language Model (LLM) agents to access dynamic, real-world tools. As MCP servers proliferate and become easy to adopt via open-source releases, understanding their security risks becomes essential for dependable production agent deployments. Recent work has developed MCP threat taxonomies, proposed mitigations, and demonstrated practical attacks. However, to the best of our knowledge, no prior study has conducted a systematic, large-scale assessment of weaknesses in open-source MCP servers. Motivated by this gap, we apply static code analysis to identify Common Weakness Enumeration (CWE) weaknesses and map them to common attack patterns and threat categories using the MITRE Common Attack Pattern Enumerations and Classifications (CAPEC) to ground risk in real-world threats. We then introduce a risk-assessment framework for the MCP landscape that combines these threats using a multi-metric scoring of likelihood and impact. Our findings show that many open-source MCP servers contain exploitable weaknesses that can compromise confidentiality, integrity, and availability, underscoring the need for secure-by-design MCP server development.

The Model Context Protocol (MCP) is a recently proposed interoperability standard that unifies how AI agents connect with external tools and data sources. By defining a set of common client-server message exchange clauses, MCP replaces fragmented integrations with a standardized, plug-and-play framework. However, to be compatible with diverse AI agents, the MCP specification relaxes many behavioral constraints into optional clauses, leading to misuse-prone SDK implementation. We identify it as a new attack surface that allows adversaries to achieve multiple attacks (e.g, silent prompt injection, DoS, etc.), named as compatibility-abusing attacks. In this work, we present the first systematic framework for analyzing this new attack surface across multi-language MCP SDKs. First, we construct a universal and language-agnostic intermediate representation (IR) generator that normalizes SDKs of different languages. Next, based on the new IR, we propose auditable static analysis with LLM-guided semantic reasoning for cross-language/clause compliance analysis. Third, by formalizing the attack semantics of the MCP clauses, we build three attack modalities and develop a modality-guided pipeline to uncover exploitable non-compliance issues.

Multi-agent systems (MAS) powered by LLMs promise adaptive, reasoning-driven enterprise workflows, yet granting agents autonomous control over tools, memory, and communication introduces attack surfaces absent from deterministic pipelines. While current research largely addresses prompt-level exploits and narrow individual vectors, it lacks a holistic architectural model for enterprise-grade security. We introduce AgenticCyOps (Securing Multi-Agentic AI Integration in Enterprise Cyber Operations), a framework built on a systematic decomposition of attack surfaces across component, coordination, and protocol layers, revealing that documented vectors consistently trace back to two integration surfaces: tool orchestration and memory management. Building on this observation, we formalize these integration surfaces as primary trust boundaries and define five defensive principles: authorized interfaces, capability scoping, verified execution, memory integrity & synchronization, and access-controlled data isolation; each aligned with established compliance standards (NIST, ISO 27001, GDPR, EU AI Act). We apply the framework to a Security Operations Center (SOC) workflow, adopting the Model Context Protocol (MCP) as the structural basis, with phase-scoped agents, consensus validation loops, and per-organization memory boundaries. Coverage analysis, attack path tracing, and trust boundary assessment confirm that the design addresses the documented attack vectors with defense-in-depth, intercepts three of four representative attack chains within the first two steps, and reduces exploitable trust boundaries by a minimum of 72% compared to a flat MAS, positioning AgenticCyOps as a foundation for securing enterprise-grade integration.

Multi-agent artificial intelligence systems or MAS are systems of autonomous agents that exercise delegated tool authority, share persistent memory, and coordinate via inter-agent communication. MAS introduces qualitatively distinct security vulnerabilities from those documented for singular AI models. Existing security and governance frameworks were not designed for these emerging attack surfaces. This study systematically characterizes the threat landscape of MAS and quantitatively evaluates 16 security frameworks for AI against it. A four-phase methodology is proposed: constructing a deep technical knowledge base of production multi-agent architectures; conducting generative AI-assisted threat modeling scoped to MAS cybersecurity risks and validated by domain experts; structuring survey plans at individual-threat granularity; and scoring each framework on a three-point scale against the cybersecurity risks. The risks were organized into 193 distinct main threat items across nine risk categories. The expected minimal average score is 2. No reviewed framework achieves majority coverage of any single category. Non-Determinism (mean score 1.231 across all 16 frameworks) and Data Leakage (1.340) are the most under-addressed domains. The OWASP Agentic Security Initiative leads overall at 65.3% coverage and in the design phase; the CDAO Generative AI Responsible AI Toolkit leads in development and operational coverage. These results provide the first empirical cross-framework comparison for MAS security and offer evidence-based guidance for framework selection. Please check back for information on the published journal version.

As multi-agent AI systems grow in complexity, the protocols connecting them constrain their capabilities. Current protocols such as A2A and MCP do not expose model-level properties as first-class primitives, ignoring properties fundamental to effective delegation: model identity, reasoning profile, quality calibration, and cost characteristics. We present the LLM Delegate Protocol (LDP), an AI-native communication protocol introducing five mechanisms: (1) rich delegate identity cards with quality hints and reasoning profiles; (2) progressive payload modes with negotiation and fallback; (3) governed sessions with persistent context; (4) structured provenance tracking confidence and verification status; (5) trust domains enforcing security boundaries at the protocol level. We implement LDP as a plugin for the JamJet agent runtime and evaluate against A2A and random baselines using local Ollama models and LLM-as-judge evaluation. Identity-aware routing achieves ~12x lower latency on easy tasks through delegate specialization, though it does not improve aggregate quality in our small delegate pool; semantic frame payloads reduce token count by 37% (p=0.031) with no observed quality loss; governed sessions eliminate 39% token overhead at 10 rounds; and noisy provenance degrades synthesis quality below the no-provenance baseline, arguing that confidence metadata is harmful without verification. Simulated analyses show architectural advantages in attack detection (96% vs. 6%) and failure recovery (100% vs. 35% completion). This paper contributes a protocol design, reference implementation, and initial evidence that AI-native protocol primitives enable more efficient and governable delegation.

Modern vision-language-model (VLM) based graphical user interface (GUI) agents are expected not only to execute actions accurately but also to respond to user instructions with low latency. While existing research on GUI-agent security mainly focuses on manipulating action correctness, the security risks related to response efficiency remain largely unexplored. In this paper, we introduce SlowBA, a novel backdoor attack that targets the responsiveness of VLM-based GUI agents. The key idea is to manipulate response latency by inducing excessively long reasoning chains under specific trigger patterns. To achieve this, we propose a two-stage reward-level backdoor injection (RBI) strategy that first aligns the long-response format and then learns trigger-aware activation through reinforcement learning. In addition, we design realistic pop-up windows as triggers that naturally appear in GUI environments, improving the stealthiness of the attack. Extensive experiments across multiple datasets and baselines demonstrate that SlowBA can significantly increase response length and latency while largely preserving task accuracy. The attack remains effective even with a small poisoning ratio and under several defense settings. These findings reveal a previously overlooked security vulnerability in GUI agents and highlight the need for defenses that consider both action correctness and response efficiency. Code can be found in https://github.com/tu-tuing/SlowBA.