Prompt Injection Defense

Educational LLM tutors face a core AI alignment challenge: they must follow user intent while preserving pedagogical constraints and safety policies. We present an evaluation methodology for prompt-injection defenses in this setting, showing that guardrail design entails explicit trade-offs among adversarial robustness, benign-task usability, and response latency. We evaluate a domain-specific multi-layer safeguard pipeline combining deterministic pattern filters, structural validation, contextual sandboxing, and session-level behavioral checks. On a controlled holdout benchmark, the pipeline reaches low bypass and false positive rates with optimized average latency - an operating point that prioritizes pedagogical usability (zero false positives) while maintaining measurable attack resistance. We provide a reproducible benchmark protocol for head-to-head comparison under identical conditions, including stratified bootstrap confidence intervals, paired McNemar significance tests, multi-seed sensitivity sweeps, and direct evaluation of Prompt Guard and NeMo Guardrails on the same split with unified instrumentation. Results expose operational trade-offs: NeMo reaches 0 percent bypass at 16.22 percent FPR and roughly 1.5s latency, while Prompt Guard yields 38.48 percent bypass with 3.60 percent FPR. The framework supports evidence-based guardrail selection for AI tutoring systems under different institutional risk and usability requirements.

Deploying large language models (LLMs) as autonomous browser agents exposes a significant attack surface in the form of Indirect Prompt Injection (IPI). Cloud-based defenses can provide strong semantic analysis, but they introduce latency and raise privacy concerns. We present the Cognitive Firewall, a three-stage split-compute architecture that distributes security checks across the client and the cloud. The system consists of a local visual Sentinel, a cloud-based Deep Planner, and a deterministic Guard that enforces execution-time policies. Across 1,000 adversarial samples, edge-only defenses fail to detect 86.9% of semantic attacks. In contrast, the full hybrid architecture reduces the overall attack success rate (ASR) to below 1% (0.88% under static evaluation and 0.67% under adaptive evaluation), while maintaining deterministic constraints on side-effecting actions. By filtering presentation-layer attacks locally, the system avoids unnecessary cloud inference and achieves an approximately 17,000x latency advantage over cloud-only baselines. These results indicate that deterministic enforcement at the execution boundary can complement probabilistic language models, and that split-compute provides a practical foundation for securing interactive LLM agents.

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.

Prompt injection poses serious security risks to real-world LLM applications, particularly autonomous agents. Although many defenses have been proposed, their robustness against adaptive attacks remains insufficiently evaluated, potentially creating a false sense of security. In this work, we propose PISmith, a reinforcement learning (RL)-based red-teaming framework that systematically assesses existing prompt-injection defenses by training an attack LLM to optimize injected prompts in a practical black-box setting, where the attacker can only query the defended LLM and observe its outputs. We find that directly applying standard GRPO to attack strong defenses leads to sub-optimal performance due to extreme reward sparsity -- most generated injected prompts are blocked by the defense, causing the policy's entropy to collapse before discovering effective attack strategies, while the rare successes cannot be learned effectively. In response, we introduce adaptive entropy regularization and dynamic advantage weighting to sustain exploration and amplify learning from scarce successes. Extensive evaluation on 13 benchmarks demonstrates that state-of-the-art prompt injection defenses remain vulnerable to adaptive attacks. We also compare PISmith with 7 baselines across static, search-based, and RL-based attack categories, showing that PISmith consistently achieves the highest attack success rates. Furthermore, PISmith achieves strong performance in agentic settings on InjecAgent and AgentDojo against both open-source and closed-source LLMs (e.g., GPT-4o-mini and GPT-5-nano). Our code is available at https://github.com/albert-y1n/PISmith.

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.

Prompt injection defenses are often framed as semantic understanding problems and delegated to increasingly large neural detectors. For the first screening layer, however, the requirements are different: the detector runs on every request and therefore must be fast, deterministic, non-promptable, and auditable. We introduce Mirror, a data-curation design pattern that organizes prompt injection corpora into matched positive and negative cells so that a classifier learns control-plane attack mechanics rather than incidental corpus shortcuts. Using 5,000 strictly curated open-source samples -- the largest corpus supportable under our public-data validity contract -- we define a 32-cell mirror topology, fill 31 of those cells with public data, train a sparse character n-gram linear SVM, compile its weights into a static Rust artifact, and obtain 95.97% recall and 92.07% F1 on a 524-case holdout at sub-millisecond latency with no external model runtime dependencies. On the same holdout, our next line of defense, a 22-million-parameter Prompt Guard~2 model reaches 44.35% recall and 59.14% F1 at 49,ms median and 324,ms p95 latency. Linear models still leave residual semantic ambiguities such as use-versus-mention for later pipeline layers, but within that scope our results show that for L1 prompt injection screening, strict data geometry can matter more than model scale.

Large language model (LLM) agents increasingly rely on external tools and retrieval systems to autonomously complete complex tasks. However, this design exposes agents to indirect prompt injection (IPI), where attacker-controlled context embedded in tool outputs or retrieved content silently steers agent actions away from user intent. Unlike prompt-based attacks, IPI unfolds over multi-turn trajectories, making malicious control difficult to disentangle from legitimate task execution. Existing inference-time defenses primarily rely on heuristic detection and conservative blocking of high-risk actions, which can prematurely terminate workflows or broadly suppress tool usage under ambiguous multi-turn scenarios. We propose AgentSentry, a novel inference-time detection and mitigation framework for tool-augmented LLM agents. To the best of our knowledge, AgentSentry is the first inference-time defense to model multi-turn IPI as a temporal causal takeover. It localizes takeover points via controlled counterfactual re-executions at tool-return boundaries and enables safe continuation through causally guided context purification that removes attack-induced deviations while preserving task-relevant evidence. We evaluate AgentSentry on the AgentDojo benchmark across four task suites, three IPI attack families, and multiple black-box LLMs. AgentSentry eliminates successful attacks and maintains strong utility under attack, achieving an average Utility Under Attack (UA) of 74.55 %, improving UA by 20.8 to 33.6 percentage points over the strongest baselines without degrading benign performance.

Large Language Models (LLMs) are widely deployed in real-world systems. Given their broader applicability, prompt engineering has become an efficient tool for resource-scarce organizations to adopt LLMs for their own purposes. At the same time, LLMs are vulnerable to prompt-based attacks. Thus, analyzing this risk has become a critical security requirement. This work evaluates prompt-injection and jailbreak vulnerability using a large, manually curated dataset across multiple open-source LLMs, including Phi, Mistral, DeepSeek-R1, Llama 3.2, Qwen, and Gemma variants. We observe significant behavioural variation across models, including refusal responses and complete silent non-responsiveness triggered by internal safety mechanisms. Furthermore, we evaluated several lightweight, inference-time defence mechanisms that operate as filters without any retraining or GPU-intensive fine-tuning. Although these defences mitigate straightforward attacks, they are consistently bypassed by long, reasoning-heavy prompts.

Large Language Model (LLM) agents are susceptible to Indirect Prompt Injection (IPI) attacks, where malicious instructions in retrieved content hijack the agent's execution. Existing defenses typically rely on strict filtering or refusal mechanisms, which suffer from a critical limitation: over-refusal, prematurely terminating valid agentic workflows. We propose ICON, a probing-to-mitigation framework that neutralizes attacks while preserving task continuity. Our key insight is that IPI attacks leave distinct over-focusing signatures in the latent space. We introduce a Latent Space Trace Prober to detect attacks based on high intensity scores. Subsequently, a Mitigating Rectifier performs surgical attention steering that selectively manipulate adversarial query key dependencies while amplifying task relevant elements to restore the LLM's functional trajectory. Extensive evaluations on multiple backbones show that ICON achieves a competitive 0.4% ASR, matching commercial grade detectors, while yielding a over 50% task utility gain. Furthermore, ICON demonstrates robust Out of Distribution(OOD) generalization and extends effectively to multi-modal agents, establishing a superior balance between security and efficiency.

Prompt injection can cause large language models (LLMs) to deviate from their intended tasks by following instructions embedded in untrusted input. We investigate whether changing a classifier's output labels improves accuracy under such attacks, using sentiment classification as a case study. Label Disguise Defense (LDD) replaces positive/negative with alternatives such as good/bad or cat/dog, introduced through demonstrations, and maps predictions back to sentiment classes. We analyze retained aggregate results for 200 IMDb reviews, nine recorded model configurations, and eight alternative label pairs under a fixed attack that always requests the incorrect class using the original labels. Averaged over four pairs per group, sentiment-associated labels outperform the other tested labels in all nine configurations, but exceed a baseline using the original labels and the same demonstration and attack conditions in only three. At least one alternative pair improves on this baseline in seven configurations, with the best pairs identified retrospectively. These results show that a consistent advantage between label groups does not imply consistent improvement over standard labels. Missing matched evaluations on unattacked reviews prevent separating changes in ordinary classification performance from changes in the effect of the attack.

Prompt injection attacks, where untrusted data contains an injected prompt to manipulate the system, have been listed as the top security threat to AI agents. By fine-tuning on simulated prompt injections, SecAlign, a leading open defense, reports LLMs with good test-time robustness and negligible benign utility drop. By scaling up training and evaluations, however, we find that SecAlign actually suffers from significant utility degradation, especially in agentic tasks where the threat of prompt injection is prominent. Motivated by this, we propose Meta-SecAlign for utility-preserving defense by (1) randomized injection position during training to avoid shortcut learning and (2) self-generated responses as high-quality in-distribution training labels. Across general knowledge, instruction following, and agentic workflows (on tool-calling and web-navigation), Meta-SecAlign maintains almost all the undefended LLM's utility while achieving better overall security than SecAlign against various static and GCG adaptive attacks. Experiments use Llama-3.1-8B, Llama-3.3-70B, Llama-4-Scout, Qwen3-4B, and Qwen3.6-27B on 6 prompt injection benchmarks including AgentDojo, InjecAgent, WASP, and SEP. Below are links for the code (https://github.com/facebookresearch/Meta_SecAlign), Meta-SecAlign-70B (https://huggingface.co/facebook/Meta-SecAlign-70B), and Meta-SecAlign-8B (https://huggingface.co/facebook/Meta-SecAlign-8B) models.

Large language models are increasingly deployed in user-facing products that handle sensitive personal data while interacting with untrusted external content. This makes them a prime target for prompt injection attacks, in which an adversary injects malicious instructions in that content to hijack the model. Fine-tuning-based defenses have been proposed to mitigate these attacks by training models to follow benign user instructions and ignore injected ones. Lacking formal guarantees, these defenses are evaluated empirically against the most effective realistic attacks available at the time, often reporting near-zero attack success rates -- giving developers, and the users who depend on them, a false sense of security and leaving open the possibility that future, more capable attacks could circumvent them and cause severe consequences in widely deployed systems. In this paper, we argue that developers should stress-test their defenses before release, and we introduce an auditing framework that does so without requiring defenders to invent the next state-of-the-art attack. Our key insight is that the fine-tuning process itself yields useful artifacts: intermediate checkpoints that serve as stepping stones for optimization-based attacks. Although these artifacts are unavailable to real attackers, we show that the vulnerabilities they expose are genuine and can be exploited by realistic adversaries. Instantiating our framework with optimization-based attacks such as GCG and BEAST, we raise attack success rates against state-of-the-art defenses from near zero to as high as 96%, revealing weaknesses missed by prior evaluations. Our results show that current evaluation practices can substantially overestimate robustness, and that auditing defenses under strong adversaries before deployment shortens the feedback loop and helps prevent attacks from causing real-world harm to users.