AI Threat AlertEvolving Jailbreaks: Automated Multi-Objective Long-Tail Attacks on Large Language Models
increases the risk of jailbreak attacks that undermine model safety alignment. While recent studies have shown that leveraging long-tail distributions can facilitate such jailbreaks, existing approaches largely rely
ECHO: A Locally-Deployable Agentic Health Assistant with Temporal Memory, Safety Guardrails, and Speech Assessment
layer intercepts all incoming queries: a rule-based layer handles explicit crisis signals and jailbreak attempts in under 1ms, while a signed graph neural network (GNN) with APPNP-style propagation
Efficient LLM Safety Evaluation through Multi-Agent Debate
Safety evaluation of large language models (LLMs) increasingly relies on LLM-as-a-judge pipelines, but strong judges can still be expensive to use at scale. We study whether structured
SkillShield: Prompt-Space Security Skills for LLM Coding Agents
settings further reduce this rate to 36.2% and 14.5%, respectively. Under two non-adaptive jailbreak families, SkillShield continues to outperform all baselines on malware generation. Across 731 benign task descriptions
Casting a SPELL: Sentence Pairing Exploration for LLM Limitation-breaking
malicious code generation. Our framework employs a time-division selection strategy that systematically constructs jailbreaking prompts by intelligently combining sentences from a prior knowledge dataset, balancing exploration of novel attack patterns
ADVERSA: Measuring Multi-Turn Guardrail Degradation and Judge Reliability in Large Language Models
Most adversarial evaluations of large language model (LLM) safety assess single prompts and report binary pass/fail outcomes, which fails to capture how safety properties evolve under sustained adversarial interaction
How Jailbreak Attacks Inform Safety Alignment: A Defender-Centric, Shapley-Based Evaluation of Jailbreak Contributions
Jailbreak attacks on large language models are usually evaluated by attacker-centric metrics such as attack success rate (ASR), yet an attack that breaks a model is not necessarily useful
Sequential Comics for Jailbreaking Multimodal Large Language Models via Structured Visual Storytelling
Multimodal large language models (MLLMs) exhibit remarkable capabilities but remain susceptible to jailbreak attacks exploiting cross-modal vulnerabilities. In this work, we introduce a novel method that leverages sequential comic
Open-Weight LLM Fine-Tuning Defenses are Susceptible to Simple Attacks
Recent defenses for safeguarding open-weight large language models (LLMs
The Laminar Flow Hypothesis: Detecting Jailbreaks via Semantic Turbulence in Large Language Models
Large Language Models (LLMs) become ubiquitous, the challenge of securing them against adversarial "jailbreaking" attacks has intensified. Current defense strategies often rely on computationally expensive external classifiers or brittle lexical
Black-Box Behavioral Distillation Breaks Safety Alignment in Medical LLMs
harmful prompt generation, verifier filtering, category-wise failure analysis, and adaptive Random Search (RS) jailbreak attacks. We also propose a layered defense system, as a prototype detector for real-time
IndicGuard: A Multilingual Safety Guard Model and Dataset for Indic Languages
Indic languages, systematically curated to capture regional harms, sensitive socio-political contexts, and adversarial jailbreaks. Leveraging this corpus, we fine-tune a 4B-parameter instruction-tuned model based on Gemma
Prompt Control-Flow Integrity: A Priority-Aware Runtime Defense Against Prompt Injection in LLM Systems
often treat prompts as flat strings and rely on ad hoc filtering or static jailbreak detection. This paper proposes Prompt Control-Flow Integrity (PCFI), a priority-aware runtime defense that
Dialect vs Demographics: Quantifying LLM Bias from Implicit Linguistic Signals vs. Explicit User Profiles
AAVE, Singlish) across various sensitive domains. Our results uncover a unique paradox in LLM safety where users achieve ``better'' performance by sounding like a demographic than by stating they belong
Refusal Evaluation in Coding LLMs and Code Agents: A Systematic Review of Thirteen Malicious-Code Prompt Corpora (2023-2025)
validated (or not) against different inter-rater reliability standards. Existing surveys treat code security, jailbreak taxonomy, or vulnerability detection as the central object and mention these corpora only in passing
OpenGuardrails: A Configurable, Unified, and Scalable Guardrails Platform for Large Language Models
such as harmful or explicit text generation, (2) model-manipulation attacks including prompt injection, jailbreaks, and code-interpreter abuse, and (3) data leakage involving sensitive or private information. Unlike prior
TraceSafe: A Systematic Assessment of LLM Guardrails on Multi-Step Tool-Calling Trajectories
hallucinations, interface inconsistencies), featuring over 1,000 unique execution instances. Our evaluation of 13 LLM-as-a-guard models and 7 specialized guardrails yields three critical findings: 1) Structural Bottleneck
VeriHarm-Bench: Measuring Harmful Success in LLM Coding Agents Verifiably
Large language model (LLM) coding agents increasingly operate with shell access, allowing them to write scripts, move files, alter permissions, and modify system configurations. However, an agent may achieve
The Workload-Router-Pool Architecture for LLM Inference Optimization: A Vision Paper from the vLLM Semantic Router Project
feedback-driven routing adaptation, hallucination detection, and hierarchical content-safety classification for privacy and jailbreak protection; (2) fleet optimization -- fleet provisioning and energy-efficiency analysis; (3) agentic and multimodal routing
ADAG: Automatically Describing Attribution Graphs
gradient effects. We then introduce a novel clustering algorithm for grouping features, and an LLM explainer--simulator setup which generates and scores natural-language explanations of the functional role