Papers

We formulate operating-system vulnerability discovery as a repeated Bayesian Stackelberg search game in which a Large Reasoning Model (LRM) orchestrator allocates analysis budget across kernel files, functions, and attack paths while external verifiers – static analyzers, fuzzers, and sanitizers – provide evidence. At each round, the orchestrator selects a target component, an analysis method, and a time budget; observes tool outputs; updates Bayesian beliefs over latent vulnerability states; and re-solves the game to minimize the strategic attacker's expected payoff. We introduce VCAO (Verifier-Centered Agentic Orchestration), a six-layer architecture comprising surface mapping, intra-kernel attack-graph construction, game-theoretic file/function ranking, parallel executor agents, cascaded verification, and a safety governor. Our DOBSS-derived MILP allocates budget optimally across heterogeneous analysis tools under resource constraints, with formal $\tilde{O}(\sqrt{T})$ regret bounds from online Stackelberg learning. Experiments on five Linux kernel subsystems – replaying 847 historical CVEs and running live discovery on upstream snapshots – show that VCAO discovers $2.7\times$ more validated vulnerabilities per unit budget than coverage-only fuzzing, $1.9\times$ more than static-analysis-only baselines, and $1.4\times$ more than non-game-theoretic multi-agent pipelines, while reducing false-positive rates reaching human reviewers by 68%. We release our simulation framework, synthetic attack-graph generator, and evaluation harness as open-source artifacts.

As generative artificial intelligence evolves, autonomous agent networks present a powerful paradigm for interactive covert communication. However, because agents dynamically update internal memories via environmental interactions, existing methods face a critical structural vulnerability: cognitive asymmetry. Conventional approaches demand strict cognitive symmetry, requiring identical sequence prefixes between the encoder and decoder. In dynamic deployments, inevitable prefix discrepancies destroy synchronization, inducing severe channel degradation. To address this core challenge of cognitive asymmetry, we propose the Asymmetric Collaborative Framework (ACF), which structurally decouples covert communication from semantic reasoning via orthogonal statistical and cognitive layers. By deploying a prefix-independent decoding paradigm governed by a shared steganographic configuration, ACF eliminates the reliance on cognitive symmetry. Evaluations on realistic memory-augmented workflows demonstrate that under severe cognitive asymmetry, symmetric baselines suffer severe channel degradation, whereas ACF uniquely excels across both semantic fidelity and covert communication. It maintains computational indistinguishability, enabling reliable secret extraction with provable error bounds, and providing robust Effective Information Capacity guarantees for modern agent networks.

Network traffic, as a key media format, is crucial for ensuring security and communications in modern internet infrastructure. While existing methods offer excellent performance, they face two key bottlenecks: (1) They fail to capture multidimensional semantics beyond unimodal sequence patterns. (2) Their black box property, i.e., providing only category labels, lacks an auditable reasoning process. We identify a key factor that existing network traffic datasets are primarily designed for classification and inherently lack rich semantic annotations, failing to generate human-readable evidence report. To address data scarcity, this paper proposes a Byte-Grounded Traffic Description (BGTD) benchmark for the first time, combining raw bytes with structured expert annotations. BGTD provides necessary behavioral features and verifiable chains of evidence for multimodal reasoning towards explainable encrypted traffic interpretation. Built upon BGTD, this paper proposes an end-to-end traffic-language representation framework (mmTraffic), a multimodal reasoning architecture bridging physical traffic encoding and semantic interpretation. In order to alleviate modality interference and generative hallucinations, mmTraffic adopts a jointly-optimized perception-cognition architecture. By incorporating a perception-centered traffic encoder and a cognition-centered LLM generator, mmTraffic achieves refined traffic interpretation with guaranteed category prediction. Extensive experiments demonstrate that mmTraffic autonomously generates high-fidelity, human-readable, and evidence-grounded traffic interpretation reports, while maintaining highly competitive classification accuracy comparing to specialized unimodal model (e.g., NetMamba). The source code is available at https://github.com/lgzhangzlg/Multimodal-Reasoning-with-LLM-for-Encrypted-Traffic-Interpretation-A-Benchmark

Ensuring reliability in adversarial settings necessitates treating privacy as a foundational component of data-driven systems. While differential privacy and cryptographic protocols offer strong guarantees, existing schemes rely on a fixed privacy budget, leading to a rigid utility-privacy trade-off that fails under heterogeneous user trust. Moreover, noise-only differential privacy preserves geometric structure, which inference attacks exploit, causing privacy leakage. We propose TADP-RME (Trust-Adaptive Differential Privacy with Reverse Manifold Embedding), a framework that enhances reliability under varying levels of user trust. It introduces an inverse trust score in the range [0,1] to adaptively modulate the privacy budget, enabling smooth transitions between utility and privacy. Additionally, Reverse Manifold Embedding applies a nonlinear transformation to disrupt local geometric relationships while preserving formal differential privacy guarantees through post-processing. Theoretical and empirical results demonstrate improved privacy-utility trade-offs, reducing attack success rates by up to 3.1 percent without significant utility degradation. The framework consistently outperforms existing methods against inference attacks, providing a unified approach for reliable learning in adversarial environments.

Linguistic steganography involves embedding secret messages within seemingly innocuous texts to enable covert communication. Provable security, which is a long-standing goal and key motivation, has been extended to language-model-based steganography. Previous provably secure approaches have achieved perfect imperceptibility, measured by zero Kullback-Leibler (KL) divergence, but at the expense of embedding capacity. In this paper, we attempt to directly use a classic entropy coding method (range coding) to achieve secure steganography, and then propose an efficient and provably secure linguistic steganographic method with a rotation mechanism. Experiments across various language models show that our method achieves around 100% entropy utilization (embedding efficiency) for embedding capacity, outperforming the existing baseline methods. Moreover, it achieves high embedding speeds (up to 1554.66 bits/s on GPT-2). The code is available at github.com/ryehr/RRC_steganography.

Talking-head generation has advanced rapidly with diffusion-based generative models, but training usually depends on centralized face-video and speech datasets, raising major privacy concerns. The problem is more acute for personalized talking-head generation, where identity-specific data are highly sensitive and often cannot be pooled across users or devices. PrivFedTalk is presented as a privacy-aware federated framework for personalized talking-head generation that combines conditional latent diffusion with parameter-efficient identity adaptation. A shared diffusion backbone is trained across clients, while each client learns lightweight LoRA identity adapters from local private audio-visual data, avoiding raw data sharing and reducing communication cost. To address heterogeneous client distributions, Identity-Stable Federated Aggregation (ISFA) weights client updates using privacy-safe scalar reliability signals computed from on-device identity consistency and temporal stability estimates. Temporal-Denoising Consistency (TDC) regularization is introduced to reduce inter-frame drift, flicker, and identity drift during federated denoising. To limit update-side privacy risk, secure aggregation and client-level differential privacy are applied to adapter updates. The implementation supports both low-memory GPU execution and multi-GPU client-parallel training on heterogeneous shared hardware. Comparative experiments on the present setup across multiple training and aggregation conditions with PrivFedTalk, FedAvg, and FedProx show stable federated optimization and successful end-to-end training and evaluation under constrained resources. The results support the feasibility of privacy-aware personalized talking-head training in federated environments, while suggesting that stronger component-wise, privacy-utility, and qualitative claims need further standardized evaluation.

Humanitarian organizations help to ensure people's livelihoods in crisis situations. Typically, multiple organizations operate in the same region. To ensure that the limited budget of these organizations can help as many people as possible, organizations perform cross-organizational deduplication to detect duplicate registrations and ensure recipients receive aid from at most one organization. Current deduplication approaches risk privacy harm to vulnerable aid recipients by sharing their data with other organizations. We analyzed the needs of humanitarian organizations to identify the requirements for privacy-friendly cross-organizational deduplication fit for real-life humanitarian missions. We present xDup, a new practical deduplication system that meets the requirements of humanitarian organizations and is two orders of magnitude faster than current solutions. xDup builds on Fuzzy PSI, and we present otFPSI, a concretely efficient Fuzzy PSI protocol for Hamming Space without input assumptions. We show that it is more efficient than existing Fuzzy PSI protocols.

The rapid expansion of the Internet of Things (IoT) and smart home ecosystems has led to a fragmented landscape of user data management across consumer electronics (CE) such as Smart TVs, gaming consoles, and set-top boxes. Current onboarding processes on these devices are characterized by high friction due to manual data entry and opaque data-sharing practices. This paper introduces the User Data Sharing System (UDSS), a platform-agnostic framework designed to facilitate secure, privacy-first PII (Personally Identifiable Information) exchange between device platforms and third-party applications. Our system implements a Contextual Scope Enforcement (CSE) mechanism that programmatically restricts data exposure based on user intent - specifically distinguishing between Sign-In and Sign-Up workflows. Unlike cloud-anchored identity standards such as FIDO2/WebAuthn, UDSS is designed for shared, device-centric CE environments where persistent user-to-device binding cannot be assumed. We further propose a tiered access model that balances developer needs with regulatory compliance (GDPR/CCPA). A proof-of-concept implementation on a reference ARMv8 Linux-based middleware demonstrates that UDSS reduces user onboarding latency by 65% and measurably reduces PII over-exposure risk through protocol-enforced data minimization. This framework provides a standardized approach to identity management in the heterogeneous CE market.

Existing red-teaming studies on GUI agents face two fundamental limitations: adversarial perturbations require white-box access unavailable in commercial deployments, while prompt injection is increasingly neutralized by stronger safety alignment. To study robustness under a more practical threat model, we propose Semantic-level UI Element Injection, a black-box red-teaming paradigm that overlays safety-aligned and harmless UI elements onto screenshots to misdirect the agent's visual grounding. Our method couples a modular Editor–Overlapper–Victim pipeline with iterative search that samples multiple candidate edits, keeps the best cumulative overlay, and adapts future prompt strategies based on previous failures. Experiments across 19 victim models spanning 8 model families show that strategic optimization substantially outperforms random injection (3.5-6.9x on the most robust victims) and transfers near-perfectly across architectures, confirming model-agnostic visual-semantic vulnerabilities. After the first successful attack, the victim still clicks the attacker-controlled icon in over 15% of subsequent independent trials versus below 1% for random injection, establishing that strategically placed icons act as persistent attractors that causally redirect grounding rather than introducing incidental clutter.

Searchable Encryption (SE) enables users to query outsourced encrypted data while preserving data confidentiality. However, most efficient schemes still leak the search pattern and access pattern, which may allow an honest-but-curious cloud server to infer query contents, user interests, or returned records from repeated searches and observed results. Existing pattern-hiding solutions mainly target keyword queries and do not naturally support Boolean range queries over encrypted spatial data. This paper presents BRASP, a searchable encryption scheme for Boolean range queries over encrypted spatial data. BRASP combines Hilbert-curve-based prefix encoding with encrypted prefix–ID and keyword–ID inverted indexes to support efficient spatial range filtering and conjunctive keyword matching. To hide the search pattern and access pattern under a dual-server setting, BRASP integrates index shuffling for encrypted keyword and prefix entries with ID-field redistribution across two non-colluding cloud servers. BRASP also supports dynamic updates and achieves forward security. We formalize the security of BRASP through confidentiality, shuffle indistinguishability, query unforgeability, and forward-security analyses, and we evaluate its performance experimentally on a real-world dataset. The results show that BRASP effectively protects query privacy while incurring relatively low computation and communication overhead. To facilitate reproducibility and further research, the source code of BRASP is publicly available at https://github.com/Egbert-Lannister/BRASP

Collaboration and information sharing empower Multi-Agent Systems (MAS) but also introduce a critical security risk known as Agent Cascading Injection (ACI). In such attacks, a compromised agent exploits inter-agent trust to propagate malicious instructions, causing cascading failures across the system. However, existing studies consider only limited attack strategies and simplified MAS settings, limiting their generalizability and comprehensive evaluation. To bridge this gap, we introduce ACIArena, a unified framework for evaluating the robustness of MAS. ACIArena offers systematic evaluation suites spanning multiple attack surfaces (i.e., external inputs, agent profiles, inter-agent messages) and attack objectives (i.e., instruction hijacking, task disruption, information exfiltration). Specifically, ACIArena establishes a unified specification that jointly supports MAS construction and attack-defense modules. It covers six widely used MAS implementations and provides a benchmark of 1,356 test cases for systematically evaluating MAS robustness. Our benchmarking results show that evaluating MAS robustness solely through topology is insufficient; robust MAS require deliberate role design and controlled interaction patterns. Moreover, defenses developed in simplified environments often fail to transfer to real-world settings; narrowly scoped defenses may even introduce new vulnerabilities. ACIArena aims to provide a solid foundation for advancing deeper exploration of MAS design principles.

Anamorphic encryption serves as a vital tool for covert communication, maintaining secrecy even during post-compromise scenarios. Particularly in the receiver-anamorphic setting, a user can shield hidden messages even when coerced into surrendering their secret keys. However, a major bottleneck in existing research is the reliance on CPA-security, leaving the construction of a generic, CCA-secure anamorphic scheme in the standard model as a persistent open challenge. To bridge this gap, we formalize the Anamorphic Key Encapsulation Mechanism (AKEM), encompassing both Public-Key (PKAKEM) and Symmetric-Key (SKAKEM) variants. We propose generic constructions for these primitives, which can be instantiated using any KEM that facilitates randomness recovery. Notably, our framework achieves strong IND-CCA (sIND-CCA) security for the covert channel. We provide a rigorous formal proof in the standard model, demonstrating resilience against a "dictator" who controls the decapsulation key. The security of our approach is anchored in the injective property of the base KEM, which ensures a unique mapping between ciphertexts and randomness. By integrating anamorphism into the KEM-DEM paradigm, our work significantly enhances the practical utility of covert channels within modern cryptographic infrastructures.

The deployment of large language models (LLMs) raises significant ethical and safety concerns. While LLM alignment techniques are adopted to improve model safety and trustworthiness, adversaries can exploit these techniques to undermine safety for malicious purposes, resulting in misalignment. Misaligned LLMs may be published on open platforms to magnify harm. To address this, additional safety alignment, referred to as realignment, is necessary before deploying untrusted third-party LLMs. This study explores the efficacy of fine-tuning methods in terms of misalignment, realignment, and the effects of their interplay. By evaluating four Supervised Fine-Tuning (SFT) and two Preference Fine-Tuning (PFT) methods across four popular safety-aligned LLMs, we reveal a mechanism asymmetry between attack and defense. While Odds Ratio Preference Optimization (ORPO) is most effective for misalignment, Direct Preference Optimization (DPO) excels in realignment, albeit at the expense of model utility. Additionally, we identify model-specific resistance, residual effects of multi-round adversarial dynamics, and other noteworthy findings. These findings highlight the need for robust safeguards and customized safety alignment strategies to mitigate potential risks in the deployment of LLMs. Our code is available at https://github.com/zhangrui4041/The-Art-of-Mis-alignment.

Existing jailbreak defense paradigms primarily rely on static detection of prompts, outputs, or internal states, often neglecting the dynamic evolution of risk during decoding. This oversight leaves risk signals embedded in decoding trajectories underutilized, constituting a critical blind spot in current defense systems. In this work, we empirically demonstrate that hidden states in critical layers during the decoding phase carry stronger and more stable risk signals than input jailbreak prompts. Specifically, the hidden representations of tokens generated during jailbreak attempts progressively approach high-risk regions in the latent space. Based on this observation, we propose TrajGuard, a training-free, decoding-time defense framework. TrajGuard aggregates hidden-state trajectories via a sliding window to quantify risk in real time, triggering a lightweight semantic adjudication only when risk within a local window persistently exceeds a threshold. This mechanism enables the immediate interruption or constraint of subsequent decoding. Extensive experiments across 12 jailbreak attacks and various open-source LLMs show that TrajGuard achieves an average defense rate of 95%. Furthermore, it reduces detection latency to 5.2 ms/token while maintaining a false positive rate below 1.5%. These results confirm that hidden-state trajectories during decoding can effectively support real-time jailbreak detection, highlighting a promising direction for defenses without model modification.

The rapid deployment of AI agents acting autonomously on behalf of human principals has outpaced the development of cryptographic protocols for establishing, bounding, and revoking human-AI trust relationships. Existing frameworks (TLS, OAuth 2.0, Macaroons) assume deterministic software and cannot address probabilistic AI agents operating continuously within variable trust boundaries. We present AITH (AI Trust Handshake), a post-quantum continuous delegation protocol. AITH introduces: (1) a Continuous Delegation Certificate signed once with ML-DSA-87 (FIPS 204, NIST Level 5), replacing per-operation signing with sub-microsecond boundary checks at 4.7M ops/sec; (2) a six-check Boundary Engine enforcing hard constraints, rate limits, and escalation triggers with zero cryptographic overhead on the critical path; (3) a push-based Revocation Protocol propagating invalidation within one second. A three-tier SHA-256 Responsibility Chain provides tamper-evident audit logging. All five security theorems are machine-verified via Tamarin Prover under the Dolev-Yao model. We validate AITH through five rounds of multi-model adversarial auditing, resolving 12 vulnerabilities across four severity layers. Simulation of 100,000 operations shows 79.5% autonomous execution, 6.1% human escalation, and 14.4% blocked.

A guard model fine-tuned on entirely benign data can lose all safety alignment – not through adversarial manipulation, but through standard domain specialization. We demonstrate this failure across three purpose-built safety classifiers – LlamaGuard, WildGuard, and Granite Guardian – deployed as protection layers in agentic AI pipelines, and show that it originates in the destruction of latent safety geometry: the structured harmful – benign representational boundary that guides classification. We extract per-layer safety subspaces via SVD on class-conditional activation differences and track how this boundary evolves under benign fine-tuning. Granite Guardian undergoes complete collapse – refusal rate drops from 85% to 0%, CKA falls to zero, and 100% of outputs become ambiguous – a severity exceeding prior findings on general-purpose LLMs, explained by the specialization hypothesis: concentrated safety representations are efficient but catastrophically brittle. To mitigate this, we propose Fisher-Weighted Safety Subspace Regularization (FW-SSR), a training-time penalty combining (i) curvature-aware direction weights derived from diagonal Fisher information and (ii) an adaptive $\lambda_t$ that scales with task-safety gradient conflict. FW-SSR recovers 75% refusal on Granite Guardian (CKA = 0.983) and reduces WildGuard's Attack Success Rate to 3.6% – below the unmodified baseline – by actively sharpening the safety subspace rather than merely anchoring it. Across all three models, structural representational geometry (CKA, Fisher score) predicts safety behavior more reliably than absolute displacement metrics, establishing geometry-based monitoring as a necessary component of guard model evaluation in agentic deployments.

We introduce Semantic Intent Fragmentation (SIF), an attack class against LLM orchestration systems where a single, legitimately phrased request causes an orchestrator to decompose a task into subtasks that are individually benign but jointly violate security policy. Current safety mechanisms operate at the subtask level, so each step clears existing classifiers – the violation only emerges at the composed plan. SIF exploits OWASP LLM06:2025 through four mechanisms: bulk scope escalation, silent data exfiltration, embedded trigger deployment, and quasi-identifier aggregation, requiring no injected content, no system modification, and no attacker interaction after the initial request. We construct a three-stage red-teaming pipeline grounded in OWASP, MITRE ATLAS, and NIST frameworks to generate realistic enterprise scenarios. Across 14 scenarios spanning financial reporting, information security, and HR analytics, a GPT-20B orchestrator produces policy-violating plans in 71% of cases (10/14) while every subtask appears benign. Three independent signals validate this: deterministic taint analysis, chain-of-thought evaluation, and a cross-model compliance judge with 0% false positives. Stronger orchestrators increase SIF success rates. Plan-level information-flow tracking combined with compliance evaluation detects all attacks before execution, showing the compositional safety gap is closable.

Precise interference detection and identification are crucial for enhancing the survivability of communication systems in non-cooperative wireless environments. While deep learning (DL) has advanced this field, existing single-task learning (STL) approaches neglect inherent task correlations. Furthermore, emerging multi-task learning (MTL) methods often lack a theoretical foundation for quantifying and modeling task relationships. To bridge this gap, we establish a theoretically grounded MTL framework for joint interference detection, modulation identification, and interference identification. First, we derive an upper bound for the weighted expected loss in MTL frameworks. This bound explicitly connects MTL performance to task similarity, quantified by the Wasserstein distance and learnable task relation coefficients. Guided by this theory, we present the adversarial multi-task interference detection and identification network (AMTIDIN), which integrates adversarial training to minimize distributional discrepancies across tasks and uses adaptive coefficients to model task correlations dynamically. Crucially, we conducted a quantitative analysis of task similarity to reveal intrinsic task relationships, specifically that modulation identification and interference identification share a substantial feature overlap distinct from interference detection. Experiments demonstrate that AMTIDIN outperforms its independently trained single-task counterparts and MTL baselines under the evaluated conditions of limited training data, short signal lengths, and low signal-to-noise ratios (SNRs)

It can be difficult for practitioners to interpret the quality of differentially private (DP) statistics due to the added noise. One method to help analysts understand the amount of error introduced by DP is to return a Randomization Interval (RI), along with the statistic. A RI is a type of confidence interval that bounds the error introduced by DP. For queries where the noise distribution depends on the input, such as the median, prior work degrades the quality of the median itself to obtain a high-quality RI. In this work, we propose PostRI, a solution to compute a RI after the median has been estimated. PostRI enables a median estimation with 14%-850% higher utility than related work, while maintaining a narrow RI.

Maximal Extractable Value, or MEV, remains a structural threat to blockchain fairness because a block producer can often observe pending transactions and unilaterally decide their ordering or inclusion. Existing mitigations hide transaction contents or outsource ordering, but they often leave two gaps unresolved. First, commitments are not authenticated by slashable identities. Second, inclusion obligations are not backed by transferable evidence that other validators can verify. This paper presents MEV ACE, a fair ordering protocol for proposer controlled ordering MEV. MEV ACE combines three mechanisms. First, it uses registered economic identities whose authentication keys are deterministically derived from the ACE GF framework and bonded on chain. Second, it uses authenticated commit and open messages with validator receipt thresholds, which make admissibility and inclusion obligations independently auditable. Third, it uses verifiable delay based randomness to determine transaction order only after the admissible commitment set is fixed. We formalize the protocol in a Byzantine fault tolerant validator model with threshold receipts and show three properties under standard assumptions: order unpredictability after the admissible set is locked, commitment authenticity under signature unforgeability, and accountable inclusion for transactions that obtain threshold commit and open receipts. Under these conditions, and when producer and user bonds exceed the one slot gain from invalid execution or selective non opening, MEV ACE removes unilateral proposer discretion over front running, sandwich attacks, and censorship against admitted transactions. The protocol remains single slot in structure, requires no threshold decryption committee, and is compatible with post quantum signature schemes such as ML DSA 44.

In event-based trust management for vehicular ad hoc networks (VANETs), vehicles that witness a road event broadcast its state, and vehicles that later witness the same event score the earlier broadcasters in feedback reports sent to a central decision unit (CDU). When the event state changes, an honest vehicle that reported the previous state receives negative feedback, as the evaluator compares an out-of-date message with the new state. This stale-witness problem is hypothesised to be a major source of unfair penalisation of honest vehicles. To address it, SAFE (Spatially-Aware Feedback Enhancement) is proposed. In SAFE, vehicles continue to record event messages after their decision and throughout the witness area, and send an updated feedback report when they leave it. SAFE was compared with the trust cascading-based emergency message dissemination model (TCEMD) in attack-free highway scenarios simulated with OMNeT++, Veins and Simulation of Urban MObility (SUMO). In the single-event scenario, the negative-feedback rate in the two update rounds after the state change decreased from 53.8% to 24.7% and from 55.6% to 8.3%, and the number of distinct honest vehicles blacklisted decreased from 20 to 6. In the multi-event scenario, the negative-feedback rate remained at or below 1.3% in SAFE, compared with up to 77.6% in TCEMD, and the share of trust evaluations ending in an untrusted label decreased from 22.7% to at most 1.7%. These gains required 2.3 to 5.0 times more feedback entries. Experiments with two decision distances showed that a shorter gap between the decision and witness distances reduced stale feedback in both schemes, supporting the stale-witness hypothesis

The Model Context Protocol (MCP) enables large language models (LLMs) to dynamically discover and invoke third-party tools, significantly expanding agent capabilities while introducing a distinct security landscape. Unlike prompt-only interactions, MCP exposes pre-execution artifacts, shared context, multi-turn workflows, and third-party supply chains to adversarial influence across independently operated components. While recent work has identified MCP-specific attacks and evaluated defenses, existing studies are largely attack-centric or benchmark-driven, providing limited guidance on where mitigation responsibility should reside within the MCP architecture. This is problematic given MCP's multi-party design and distributed trust boundaries. We present a defense-placement-oriented security analysis of MCP, introducing a layer-aligned taxonomy that organizes attacks by the architectural component responsible for enforcement. Threats are mapped across six MCP layers, and primary and secondary defense points are identified to support principled defense-in-depth reasoning under adversaries controlling tools, servers, or ecosystem components. A structured mapping of existing academic and industry defenses onto this framework reveals uneven and predominantly tool-centric protection, with persistent gaps at the host orchestration, transport, and supply-chain layers. These findings suggest that many MCP security weaknesses stem from architectural misalignment rather than isolated implementation flaws.

We present a constructive proof that a single C program, the Vulnerability Factory, admits a countably infinite set of distinct, independently CVE-assignable software vulnerabilities. We formalise the argument using elementary set theory, verify it against MITRE's CVE Numbering Authority counting rules, sketch a model-checking analysis that corroborates unbounded vulnerability generation, and provide a Turing-machine characterisation that situates the result within classical computability theory. We then contextualise this result within the long-running debate on whether undiscovered vulnerabilities in software are dense or sparse, and introduce the concept of vulnerability abundance: a quantitative analogy to chemical elemental abundance that describes the proportional distribution of vulnerability classes across the global software corpus. Because different programming languages render different vulnerability classes possible or impossible, and because language popularity shifts over time, vulnerability abundance is neither static nor uniform. Crucially, we distinguish between infinite vulnerabilities and the far smaller set of exploits: empirical evidence suggests that fewer than 6% of published CVEs are ever exploited in the wild, and that exploitation frequency depends not only on vulnerability abundance but on the market share of the affected software. We argue that measuring vulnerability abundance, and its interaction with software deployment, has practical value for both vulnerability prevention and cyber-risk analysis. We conclude that if one programme can harbour infinitely many vulnerabilities, the set of all software vulnerabilities is necessarily infinite, and we suggest the Vulnerability Factory may serve as a reusable proof artifact, a foundational `test object',for future formal results in vulnerability theory.

Large language models (LLMs) increasingly rely on external tools to perform time-sensitive tasks and real-world actions. While tool integration expands LLM capabilities, it also introduces a new prompt-injection attack surface: tool poisoning attacks (TPAs). Attackers manipulate tool descriptions by embedding malicious instructions (explicit TPAs) or misleading claims (implicit TPAs) to influence model behavior and tool selection. Existing defenses mainly detect anomalous instructions and remain ineffective against implicit TPAs. In this paper, we present TRUSTDESC, the first framework for preventing tool poisoning by automatically generating trusted tool descriptions from implementations. TRUSTDESC derives implementation-faithful descriptions through a three-stage pipeline. SliceMin performs reachability-aware static analysis and LLM-guided debloating to extract minimal tool-relevant code slices. DescGen synthesizes descriptions from these slices while mitigating misleading or adversarial code artifacts. DynVer refines descriptions through dynamic verification by executing synthesized tasks and validating behavioral claims. We evaluate TRUSTDESC on 52 real-world tools across multiple tool ecosystems. Results show that TRUSTDESC produces accurate tool descriptions that improve task completion rates while mitigating implicit TPAs at their root, with minimal time and monetary overhead.

Event-based trust management in vehicular ad hoc networks is vulnerable to strategic insider attackers who report trivial events correctly to build reputation, falsify reports on safety-critical events, and corrupt the trust feedback they provide about other vehicles. This paper proposes IPEK, an event-criticality-aware evidential trust management framework. IPEK combines an asymmetric local trust that scales rewards and penalties with event severity and location criticality, a reporter-credibility test that discounts feedback deviating systematically from the consensus of witnesses, and a Dempster-Shafer global trust that fuses feedback with Yager's rule, historical combination and risk accentuation. Closed-form properties quantify how quickly trust is lost and regained and show why Yager's rule keeps the global state revisable. IPEK is evaluated in simulation against a trust cascading-based dissemination scheme and the beta reputation system under eight attacks, with severity-, location-, on-off- and continuously triggered attackers falsifying their reports alone or also their feedback, at attacker ratios from 10% to 40%. Threshold-free separation measures are complemented by closed-loop revocation with frozen thresholds. With up to 25% attackers, IPEK achieves a significantly higher F1-score than both baselines under all attacks, with false positive rates of 0.9-15.5% against 7.8-49.7% and 26.0-41.4%, and it remains robust up to 40% attackers when only reports are falsified. Under fully colluding attackers that also corrupt feedback, all methods break down beyond about one third of attackers; this envelope and a longer detection delay are reported as limitations.

Epidemiologic studies of infectious diseases often rely on models of contact networks to capture the complex interactions that govern disease spread, and ongoing projects aim to vastly increase the scale at which such data can be collected. However, contact networks may include sensitive information, such as sexual relationships or drug use behavior. Protecting individual privacy while maintaining the scientific usefulness of the data is crucial. We propose a privacy-preserving pipeline for disease spread simulation studies based on a sensitive network that integrates differential privacy (DP) with statistical network models such as stochastic block models (SBMs) and exponential random graph models (ERGMs). Our pipeline comprises three steps: (1) compute network summary statistics using node-level DP (which corresponds to protecting individuals' contributions); (2) fit a statistical model, like an ERGM, using these summaries, which allows generating synthetic networks reflecting the structure of the original network; and (3) simulate disease spread on the synthetic networks using an agent-based model. We evaluate the effectiveness of our approach using a simple Susceptible-Infected-Susceptible (SIS) disease model under multiple configurations. We compare both numerical results, such as simulated disease incidence and prevalence, as well as qualitative conclusions such as intervention effect size, on networks generated with and without differential privacy constraints. Our experiments are based on egocentric sexual network data from the ARTNet study (a survey about HIV-related behaviors). Our results show that the noise added for privacy is small relative to other sources of error (sampling and model misspecification). This suggests that, in principle, curators of such sensitive data can provide valuable epidemiologic insights while protecting privacy.

Large language models (LLMs) have emerged as a powerful tool for synthetic data generation. A particularly important use case is producing synthetic replicas of private text, which requires carefully balancing privacy and utility. We propose Realistic and Privacy-Preserving Synthetic Data Generation (RPSG), which uses private seeds and integrates privacy-preserving strategies, including a formal differential privacy (DP) mechanism in the candidate selection, to generate realistic synthetic data. Comprehensive experiments against state-of-the-art private synthetic data generation methods demonstrate that RPSG achieves high fidelity to private data while providing strong privacy protection.

Retrieval-Augmented Generation (RAG) significantly enhances Large Language Models (LLMs), but simultaneously exposes a critical vulnerability to knowledge poisoning attacks. Existing attack methods like PoisonedRAG remain detectable due to coarse-grained separate-and-concatenate strategies. To bridge this gap, we propose RefineRAG, a novel framework that treats poisoning as a holistic word-level refinement problem. It operates in two stages: Macro Generation produces toxic seeds guaranteed to induce target answers, while Micro Refinement employs a retriever-in-the-loop optimization to maximize retrieval priority without compromising naturalness. Evaluations on NQ and MSMARCO demonstrate that RefineRAG achieves state-of-the-art effectiveness, securing a 90% Attack Success Rate on NQ, while registering the lowest grammar errors and repetition rates among all baselines. Crucially, our proxy-optimized attacks successfully transfer to black-box victim systems, highlighting a severe practical threat.

With the widespread application of artificial intelligence technologies in face recognition and other fields, data privacy security issues have received extensive attention, especially the right to be forgotten emphasized by numerous privacy protection laws. Existing technologies have proposed various unlearning methods, but they may inadvertently leak the categories of unlearned data. This paper focuses on the category unlearning scenario, analyzes the potential problems of category leakage of unlearned data in multiple scenarios, and proposes four attack methods from the perspectives of model parameters and model inversion based on attackers with different knowledge backgrounds. At the level of model parameters, we construct discriminative features by computing either dot products or vector differences between the parameters of the target model and those of auxiliary models trained on subsets of retained data and unrelated data, respectively. These features are then processed via k-means clustering, Youden's Index, and decision tree algorithms to achieve accurate identification of the forgotten class. In the model inversion domain, we design a gradient optimization-based white-box attack and a genetic algorithm-based black-box attack to reconstruct class-prototypical samples. The prediction profiles of these synthesized samples are subsequently analyzed using a threshold criterion and an information entropy criterion to infer the forgotten class. We evaluate the proposed attacks on four standard datasets against five state-of-the-art unlearning algorithms, providing a detailed analysis of the strengths and limitations of each method. Experimental results demonstrate that our approach can effectively infer the classes forgotten by the target model.

Operating LEO mega-constellations requires translating high-level operator intents ("reroute financial traffic away from polar links under 80 ms") into low-level routing constraints – a task that demands both natural language understanding and network-domain expertise. We present an end-to-end system comprising three components: (1) a GNN cost-to-go router that distills Dijkstra-quality routing into a 152K-parameter graph attention network achieving 99.8% packet delivery ratio with 17x inference speedup; (2) an LLM intent compiler that converts natural language to a typed constraint intermediate representation using few-shot prompting with a verifier-feedback repair loop, achieving 98.4% compilation rate and 87.6% full semantic match on feasible intents in a 240-intent benchmark (193 feasible, 47 infeasible); and (3) an 8-pass deterministic validator with constructive feasibility certification that achieves 0% unsafe acceptance on all 47 infeasible intents (30 labeled + 17 discovered by Pass 8), with 100% corruption detection across 240 structural corruption tests and 100% on 15 targeted adversarial attacks. End-to-end evaluation across four constrained routing scenarios confirms zero constraint violations with both routers. We further demonstrate that apparent performance gaps in polar-avoidance scenarios are largely explained by topological reachability ceilings rather than routing quality, and that the LLM compiler outperforms a rule-based baseline by 46.2 percentage points on compositional intents. Our system bridges the semantic gap between operator intent and network configuration while maintaining the safety guarantees required for operational deployment.