Prompt injection: research

Journal articles, academic blogs, think-tank reports and papers in this section, newest first. Back to the news.

arXiv (research papers) Roshan Sood SI

COPA: Continual Preference Optimization for Adaptive Prompt Injection Defense

LLMs remain vulnerable to prompt injection attacks, where adversarial instructions embedded in user inputs or external content manipulate model behavior and bypass safeguards. Existing defenses are predominantly static, relying on fixed alignment objectives or attack-specific filtering mechanisms that require redesign as new attack strategies emerge. While recent lifelong alignment methods address shifting user preferences, they do not account for adaptive adversaries that…

Read at arXiv (research papers)

arXiv (research papers) Sujin Chen SI

MobileWorldSafety: Benchmarking GUI Agent Safety Against Environmental Injection Attacks in Android Apps

LLM-powered GUI agents that autonomously operate smartphones are rapidly transitioning from research prototypes to early real-world deployment. However, because these agents routinely process untrusted environmental content, they are highly vulnerable to environmental injection attacks, which include indirect prompt injections and adversarial instructions. Such attacks can manipulate the behavior of agents without user awareness through diverse channels encountered in…

Read at arXiv (research papers)

arXiv (research papers) Rabimba Karanjai Larry SI

PACE: Policy-Attested Contract Execution for Safe AI Agents in Decentralized Finance

Autonomous AI agents are emerging as interfaces for decentralized finance (DeFi) actions such as swaps, lending operations, and yield management. Because these agents rely on large language models (LLMs) to plan transactions, they inherit the LLM's susceptibility to prompt injection and lack of mechanisms to bind a verifier's approval to the exact transaction ultimately submitted on-chain. We present PACE (Policy-Attested Contract Execution), a transaction-level…

Read at arXiv (research papers)

arXiv (research papers) Jiawei Liu SI

Security of Foundation-Model-Powered Embodied Agents: Attack Surfaces, Attacks, Defenses, and Evaluation

Foundation models are increasingly used for perception, reasoning, planning, and action generation in embodied agents, creating security risks that can propagate from digital inputs to physical behavior. Existing surveys often organize threats by mechanisms such as jailbreaks, prompt injection, backdoors, poisoning, or adversarial examples, but these categories do not consistently identify where an adversary first enters the embodied control loop. We present a…

Read at arXiv (research papers)

arXiv (research papers) Zonghao Ying SI

Security Assessment of DeepSeek Harness with A.I.G: Evaluating Resistance to Indirect Prompt Injection

We assess indirect prompt injection in DeepSeek Harness (DSH), using AI-Infra-Guard (A.I.G) to construct tests, deliver controlled taint, execute DSH, collect traces, and judge outcomes. The study covers 14,560 controlled executions over 16 indirect-content channels, text and file carrier modes, 35 payload objectives, one unmodified baseline, and 12 attack methods. The experiment preserves DSH's agent loop, tool registry, model adapter, and session-event path; source tools…

Read at arXiv (research papers)

arXiv (research papers) Jun He SI

Agent-Native Telemetry: Verifiable State-Delta Evidence for Autonomous Operations

Operational telemetry is predominantly engineered for human reading: systems repeatedly serialize verbose prose, static keys, and redundant context across billions of log lines. As autonomous AI agents become primary operational consumers, feeding them traditional logs wastes scarce context capacity parsing lexical syntax rather than reasoning over system state changes -- all while lacking cryptographic guarantees of provenance or collection completeness. This paper…

Read at arXiv (research papers)

arXiv (research papers) Xabier Muruaga SI

Bounded Agents: Delegation Security for Multi-Agent AI Systems

LLM-based agents can act on behalf of a user to access cloud services, call tools, or invoke agents. At session start, the agent's permissions are set but remain static, and each request is evaluated independently, without considering prior actions. Within its permissions, an agent may act contrary to the delegated task, combine individually permitted actions into a prohibited outcome, or delegate authority to a sub-agent without limiting it. A prompt injection poses a risk…

Read at arXiv (research papers)

arXiv (research papers) Madhusudhanan G SI

Visible Reasoning and Indirect Prompt-Injection Monitorability Across English, Tamil, and Tanglish

Chain-of-thought monitoring is a potentially useful safety signal, but its reliability across languages and behavioral settings remains uncertain. In a small case study of eight manually verified synthetic scenarios, one model, one annotator, and one deterministic generation seed, I study API-visible reasoning during indirect prompt injection in Sarvam-105B across English, Tamil, and Tanglish. A four scenario pilot found 5/12 injected attack successes without reasoning and…

Read at arXiv (research papers)

arXiv (research papers) Md Fazley Rafy SI

TwinGridShield: Consequence-Aware Runtime Authorization for LLM Grid-Agent Actions

Large language model (LLM)-assisted energy-management tools can translate natural-language context into structured grid commands, but syntactic validity does not imply physical admissibility. This paper presents TwinGridShield, a model-independent runtime authorization layer that evaluates each proposed action in a deterministic network twin before release. The prototype checks connectivity, branch-flow, generator, and load-shedding invariants and records each decision in a…

Read at arXiv (research papers)

arXiv (research papers) Anees Ahmed Mahaboob Ali SI

BioFirewall: A genome-writing-native governance layer for design-stage biosecurity screening of agentic AI

AI design tools now plan genome-scale edits, and agentic systems execute those plans with progressively less human oversight. Biosecurity controls are limited to two points: refusal guardrails at the foundation model and sequence-identity screening at the synthesiser. The design stage between them, where the plan is specified, is governed by recommendations, not by any deployed system. We present BioFirewall, a rule-governed middleware that intercepts a genome-writing plan…

Read at arXiv (research papers)

arXiv (research papers) Alexandre G. R. Day SI

Workspace Topology as an Attack Vector in Agentic Coding Assistants

Agentic coding assistants are finding widespread use, not just in new code development but in quickly ingesting and leveraging third-party code. This opens up a risk of malicious code being ingested as these coding tools operate with broad filesystem access inside developer workspaces. In this paper, we extensively study the impact of different dimensions of a novel attack surface we term workspace topology -- defined via directory depth, codebase modularity, in-file…

Read at arXiv (research papers)

arXiv (research papers) Van Tran SI

Rethinking Agent Security as a Networking Problem

AI agents are rapidly becoming more capable and widely deployed, promising substantial gains in productivity and enabling new classes of applications. However, their growing autonomy also introduces significant privacy and security risks. Existing defenses are predominantly agent-centric, relying on the agent itself to detect threats and enforce privacy and security policies. This approach is fundamentally limited because it entrusts policy enforcement to AI agents whose…

Read at arXiv (research papers)

arXiv (research papers) Yutao Mou SI

ToolHazard: Scaling Adversarial Environments for Security Evaluation and Alignment of LLM-based Agents

Large language model (LLM) agents integrated with external tools are vulnerable to indirect prompt injections embedded in environmental states. However, existing studies largely rely on manually implemented or reused environments, stochastic LLM-based tool simulation, and predefined injection locations, limiting scalable security research across broader domains. To bridge this gap, we propose **ToolHazard**, a scalable adversarial environment synthesis framework that reduces…

Read at arXiv (research papers)

arXiv (research papers) Md Jafrin Hossain SI

On Understanding, Identifying, and Mitigating Vulnerabilities in Agentic Large Language Models

Large Language Models (LLMs) have undergone a shift from stateless conversational interfaces to autonomous agents capable of multi-step planning, tool invocation, code execution, and maintaining persistent memory. When these agents operate with real-world privileges---calling APIs, modifying files, and querying databases---a compromised reasoning step can trigger unauthorized data access, irreversible state changes, or cascading failures, yet the security research community…

Read at arXiv (research papers)

arXiv (research papers) Spiros Tsigkopoulos SI

From Prompt Injection to Web Exploitation: Revisiting Classic Vulnerabilities in LLM-Integrated Applications

Large Language Models are increasingly integrated into web applications through chatbots, tool-calling pipelines, and agentic workflows. In these systems, user input may influence not only generated text, but also backend actions such as database queries, HTTP requests, file operations, template rendering, or API calls. This paper introduces LLM-mediated web attacks, a class of attacks in which attacker-controlled input is transformed by an LLM-integrated application and…

Read at arXiv (research papers)

arXiv (research papers) Jordan Pettyjohn SI

Interpreting Language Model Hidden States at Scale

Lens methods interpret large language models (LLMs) by mapping intermediate activations to the output vocabulary, revealing how next-token predictions develop through the network. Trained lenses remain expensive: affine-translator parameters grow quadratically with model width, while exact, full-vocabulary Kullback--Leibler (KL) training dominates memory. Consequently, prior trained lenses have been applied to models of at most 20B parameters and remain tied to particular…

Read at arXiv (research papers)

arXiv (research papers) Alexander Panfilov SI

Stealing Reasoning Traces from Proprietary LLM APIs

Leading large language model providers now conceal their models' step-by-step reasoning, or chain-of-thought, to protect intellectual property and limit information leakage. Rather than storing these traces server-side, providers return them to the client as blocks of encrypted text, which the client passes back with each subsequent request. Building on prior research, we identify an architectural vulnerability: these encrypted blocks are fully compatible and interchangeable…

Read at arXiv (research papers)

arXiv (research papers) Sam Siavoshian SI

ChronoState: Hidden Elapsed-Time Conditioning for Temporal-State Action Selection in Frozen-Backbone Language Models

Temporal decisions in language-model systems often depend on both symbolic task state and elapsed wall-clock time, such as cache expiration, job completion, quota resets, deadlines, or stale sessions. We study whether elapsed time can be supplied as a non-token, system-side scalar and composed with visible symbolic state by a frozen-backbone language model. We introduce ChronoState, a compositional temporal-state benchmark in which symbolic state appears in the prompt…

Read at arXiv (research papers)

arXiv (research papers) Lier Jin SI

Same Question, Different Answer? Measuring and Mitigating Prompt Privilege for Equitable AI Access

Large language models (LLMs) are increasingly integrated into healthcare, education, public services, and everyday decision making. They should provide comparable assistance regardless of a user's literacy, communication style, or prompt-engineering expertise. However, existing research on prompt robustness primarily focuses on adversarial attacks, prompt injection, and prompt optimization, while overlooking whether semantically equivalent requests receive different…

Read at arXiv (research papers)

arXiv (research papers) Rahul Deivasigamani SI

Not an A11y: How Android Accessibility Exposes Mobile AI Agents to Indirect Prompt Injection

The rise of autonomous AI agents represents a major paradigm shift in how users interact with mobile devices. Frameworks such as MobileRun and Mobile-Use can autonomously navigate Android applications and execute complex multi-step tasks. To interpret user interfaces, these frameworks rely primarily on Android accessibility (A11y) trees and secondarily on visual screenshots. In this paper, we demonstrate that this architectural dependence on unsanitized accessibility…

Read at arXiv (research papers)

arXiv (research papers) Sihan Hou SI

Toward Metacognitive One-Shot Indirect Prompt Injection: Strategy Abstraction Via Outcome-Conditioned Reflection

Tool-using large language model (LLM) agents are vulnerable to indirect prompt injection (IPI), in which malicious instructions embedded in external observations manipulate subsequent agent decisions and actions. Most existing adaptive attacks rely on repeatedly querying and refining against the target agent, whereas realistic attackers may have only a single opportunity to interact with an unknown target agent. We propose SAVOR (Strategy Abstraction Via Outcome-Conditioned…

Read at arXiv (research papers)

arXiv (research papers) Xinze Chen SI

SkillsMetric: Mapping the Detection Boundary of Static Analysis for Malicious Agent Skills

Agent Skills---structured packages of instructions and scripts that augment LLM-based agents---are rapidly proliferating, yet their security properties remain under-explored. We present \textsc{SkillsMetric}, a five-stage static analysis framework that scores skill packages along pattern density, statistical anomaly, dataflow taint, import anomaly, and capability mismatch dimensions. We construct an adversarial evaluation dataset of 2{,}266 skills spanning 16~attack types…

Read at arXiv (research papers)

arXiv (research papers) Kaysarul Anas Apurba SI

Defending Retrieval-Augmented Intrusion Detection Against Knowledge Poisoning and Prompt Injection

Retrieval-Augmented Generation (RAG) enables large language models to classify network flows and generate human-readable incident reports by retrieving semantically similar historical traffic from a vector knowledge base. However, the retrieval layer introduces vulnerabilities to knowledge poisoning and prompt-injection attacks. We present RAG-IDS, a three-tier multi-agent intrusion detection framework with a retrieval-boundary defense combining soft trust scoring…

Read at arXiv (research papers)

arXiv (research papers) Laiqiao Qin SI

BASIS: Breach-Aware Selective Prompt Injection Shielding with Prefill Attention Probes

Prompt injection is a critical security threat in large language model (LLM) applications, where attackers hijack model behavior by embedding malicious instructions in user or external data. Existing detection methods only detect the presence of injection and refuse to respond upon detection, overlooking the fact that for many modern aligned models, well-crafted instructions can resist most injection attacks. This means that the injection robustness varies significantly…

Read at arXiv (research papers)

arXiv (research papers) Jeremy McHugh SI

The Anatomy of a Prompt Injection: A Component Model for Structured Analysis

Four years after prompt injection was first identified in 2022, attacks are still predominantly documented as verbatim strings rather than structured exploits, despite advancing agent capabilities and threat actors embedding injections to subvert AI-assisted security analysis. This paper formalizes the structure of prompt-injection artifacts, enabling defenders, red teamers, and cyber threat intelligence (CTI) teams to label, compare, and mutate attacks without relying on…

Read at arXiv (research papers)

arXiv (research papers) Aditya Katkar SI

NiyamAI - An Intent-Bound AI Agent with Cryptographically Verifiable Guardrails using Zero-Knowledge Proofs

Autonomous LLM agents with tool execution capabilities introduce severe security risks through prompt injection, goal hijacking, and unauthorized action invocation. Existing guardrails rely on unverified, host local software filters system prompts, semantic classifiers, policy engines that share the execution environment of the untrusted agent, offering no guarantee to an external observer that a safety policy was correctly evaluated. A compromised host produces no evidence…

Read at arXiv (research papers)

arXiv (research papers) Zhuoxin Zhan SI

StepJack: Benchmarking Computer-Use Agent Safety Against Multi-Step Indirect Prompt Injection

Computer-use agents (CUAs) face a growing threat from indirect prompt injection, where adversarial instructions are planted in the environment such as web pages. In this paper, we introduce multi-step indirect prompt injection, a new attack class against CUAs in which the adversarial goal is decomposed into multiple innocuous-looking sub-steps and distributed across a chain of pages referenced along the agent's navigation path. We develop a pipeline to automatically…

Read at arXiv (research papers)

arXiv (research papers) Leo Sambrook SI

Hardware Keystores for AI Agent Signing Workflows: A Zero-Trust MCP Enforcement Architecture

AI agents increasingly sign Git commits, certify documents, and attest release artifacts on behalf of their operators, using private keys that live in software-accessible locations (plaintext files, environment variables, container memory) readable by any process the agent can reach. A widely deployed agent framework recently leaked its keys this way to a single email injection. Hardware keystores (HSM, TPM, smart card) keep the key on-device, but exposing the keystore as a…

Read at arXiv (research papers)

arXiv (research papers) S. M . Bhagya P. Samarakoon SI

Hijacking Robots with a Piece of Paper: A Systematic Study of Physical Prompt Injection in VLM-Controlled Robots

Vision-Language Models (VLMs) are increasingly deployed as planners in robotic systems, where they translate natural-language commands into executable actions grounded in visual scene understanding. This tight coupling between perception and instruction-following introduces a new attack surface: adversarial text placed within the robot's visual field can act as an indirect prompt injection into the VLM's reasoning stack. We present a systematic study of physical prompt…

Read at arXiv (research papers)

arXiv (research papers) He Zhang SI

PromptShield Home: Ambient Multimodal Prompt Injection Defense for Smart-Home Agents

Smart-home assistants increasingly use multimodal large language models (MLLMs) that perceive video and audio directly. This raises a safety question specific to the home: can the agent tell a genuine user command from ambient or externally-sourced content, television speech, on-screen text, or an overheard conversation, that merely looks like a command? We introduce PromptShield-Home, a pilot benchmark of realistic smart-home scenarios spanning addressee ambiguity…

Read at arXiv (research papers)

Older stories