AI Threat AlertTrusted Credentials, Untrusted Behavior: Benchmarking LLM-Agent Security in High-Performance Computing
hijacked authorized agent problem. Existing agent-security studies explain relevant mechanisms, such as indirect prompt injection and tool misuse, but generally evaluate them in web, enterprise, or personal-assistant settings
Adaptive Adversaries: A Multi-Turn, Multi-LLM Benchmark for LLM Agent Security
based agents process external content, exposing them to prompt injection and multi-turn manipulation. Most safety benchmarks evaluate defenders against fixed attack pools collected before evaluation, single-turn or multi
Persistent Sparse Autoencoders: Learning Feature Timescales in Language Models
features concentrate topic-level information in a persistent state. Moreover, as shown in a prompt-injection monitoring case study, slow features preserve detection signals and remain causally effective over long
Agent Skill Security: Threat Models, Attacks, Defenses, and Evaluation
packaged, shared, and reused across diverse applications. However, existing security research primarily focuses on prompt injection and runtime execution, leaving security risks throughout the broader skill lifecycle largely unexplored
How Agents Ask for Permission: User Permissions for AI Agents, from Interfaces to Enforcement
agents gain prevalance, users are increasingly exposed to the risks such systems entail. Prompt injection attacks, as well as hallucination, can cause agents to leak private information to third parties
Self-Evolving Agent Harnesses via Gated Semantic Quality-Diversity
much by the harness around its model as by the frozen model itself: its prompts, injected knowledge, runtime control, and configuration. In deployment the harness is often the only lever
Isolation as a First-Class Principle for LLM-Agent System Safety: Concepts, Taxonomy, Challenges and Future Directions
types, applications, and benchmarks. This makes it hard to explain why failures such as prompt injection, tool misuse, and memory poisoning often share the same structural cause, and how they
Distributed Denial of Science: How Indirect Data Poisoning of AI Systems Can Industrialize Scientific Fraud
only 6.0%. The attack requires no topic-specific trigger-words, agent access, indirect prompt injection, or fabricated papers, only the open data ecosystem and misleading metadata. To mitigate the attacks
Multi-Agent Firewall Architecture for Privacy Protection of Sensitive Data in Interactions with Language Models
proprietary code leakage prevention, and extensible components designed for future security enhancements such as prompt injection evasion. The framework's layered architecture enables deployment across heterogeneous environments, allowing organizations
FedMark-FM: Auditable, Risk-Adjusted Data Markets for Federated Foundation-Model Adaptation
retrieval, held-out generator-backed RAG, and trained PEFT/LoRA tracks. Under a held-out prompt-injection poisoner, FedMark-FM improves downstream accuracy by 7.5-8.1 points over volume, leave
Untrusted Content Masking for Web Agents with Security Guarantees
Defenses that provide security guarantees against prompt injection attacks rely on strict isolation between trusted instructions and untrusted data. In text-based environments such as tool-use APIs, this separation
Biological Motifs for Agentic Control
agentic architectures are often constructed ad-hoc, prone to hallucination cascades, infinite loops, and prompt injection attacks. This paper argues that many of these failure modes can be analyzed using
Distributed Attacks in Persistent-State AI Control
codebase persisting across sessions. This persistence creates a new attack surface: a misaligned or prompt-injected agent can distribute attacks across pull requests (PRs) and time its payload
Adversarial Pragmatics for AI Safety Evaluation: A Benchmark for Instruction Conflict, Embedded Commands, and Policy Ambiguity
methodology into a practical tool for validating safety evals, LLM judges, gold-set construction, prompt-injection tests, and safety documentation
Understanding and Evaluating Claw-like Agent Security Through a Computer-Systems Lens
surfaces (Skill Supply-Chain Integrity, Persistent State Exploitation, Cross-Boundary Data Flow, and Indirect Prompt Injection), executed in containerized replicas of real agent platforms with canary-marked credentials and evaluated
IHDec: Divergence-Steered Contrastive Decoding for Securing Multi-Turn Instruction Hierarchies
turn conflicts while fully preserving general response quality. Furthermore, IHDec strengthens safety against adversarial prompt injections and exhibits a robust scaling synergy with larger models. The Code is available
Agents That Know Too Much: A Data-Centric Survey of Privacy in LLM Agents
active but scattered across retrieval-augmented generation, text-to-SQL interfaces, agent memory, prompt injection, access control, and contextual privacy. This survey brings that work together: we taxonomize the data
Beyond Feedforward Networks: Reentry Neural Systems as the Fundamental Basis of Subjecthood and Intrinsic Safety of Next-Generation AGI
textual D-vector in the architecture itself, making them immune to reinterpretation and prompt injection. We present the S-measure -- a polynomial-time [O(N^3)] computable alternative to Tononi
AI Snitches Get Glitches: Towards Evading Agentic Surveillance
they also report the attempts to surveil users to the government. Finally, we repurpose prompt injections for evading surveillance and develop three evasion techniques that hide from, deceive, or induce
PARSE: Provenance-Aware Retrieval Sanitization for Professional Domain LLM Agents
Prompt injection defenses evaluated on synthetic benchmarks do not generalize to real enterprise documents, which are longer, denser, and interleave legitimate authority language with factual content. We demonstrate this