Research
Verified developments, reporting, and emerging signals in research.
BiShield-TEE: On the (In-)Security of Unilateral Weight Obfuscation in On-Device TEE-Shielded LLM Partition
Poisoning the Neighborhood: A Topology-Aware Topic-Level Attack on Large-Scale HNSW-based RAG Systems
QuasiLeak-Bench: Evaluating LLM-Based Privacy Detectors on Policy-Conditioned Quasi-Evidence Group Recovery
Combining LLM and Dynamic Analysis for Universal Testing of Static Application Security Testing Tools
From One Query to Many: A Critical Survey of Privacy Accounting in RAG
Stronger Judges Do Not See More: Deployment-Integrity Blindness in LLM Evaluation
MMA-SafetyBench: A Benchmark for Multimodal Agent Safety Evaluation
AgentCIScope: Grounded Contextual Integrity Evaluation for LLM Agents
Rethinking Quality in Large Language Model Routing: Security-Conditioned Model-and-Budget Selection
FedGraph: Defending Federated Large Language Model Fine-Tuning Against Backdoor Attacks via Graph-Based Aggregation
Pin It on Anyone: Certified Targeted Displacement of Multi-Bit LLM Watermarks Without the Key
Holding Agents Accountable with Verifiable Agent Insurance
CompanionSafe: From Multi-turn Safety Evaluation to Guardrails for AI Companions
Cost-Efficient Privacy Enforcement for Email-Generating LLM Agents under Strategic Information Requests
Anthropic takes risky AI tests offline after models breach govt sites
EviRepair: Evidence-Driven Agents for Reliable Vulnerability Repair
Understanding and Mitigating Cross-Session Attacks via Persistent Artifacts in Agent Skill Composition
ReSI: Recursive Safety Improvement toward Resistant and Resilient AI
Poster: A Preliminary Study of LLM Distillation Inference
Protecting CPU AI On Edge TEEs: WebAssembly's Promise and Practical Challenges
Anytime-valid detection of LLM weight exfiltration
SoK: Are LLMs Reliable at Source Code Recovery? A Taxonomy and Empirical Evaluation
Claude Haiku 5.5
On the Reliability of LLM-Based Vulnerability Patching Benchmarks
Efficient Provably Private Classification with a Tabular Foundation Model
Defensive Sufficiency in a Stackelberg Model of AI Security
Package Hallucination Attacks on Coding Agents through Prompt Injection in Rule Files
BARE-AI: Bit-Flip Attack Resilience in AI Hardware through Built-in Performance Monitors
RAG-PIBench: A Leakage-Aware Benchmark for Prompt-Injection Detection in Trustworthy RAG Systems
Case-Level Verification in Scanner-LLM Cascades: Overcoming the Alert Aggregation Bottleneck to Expand the FRR-TPR Trade-off Space
Learning from Failures: A Failure-Driven Prompt Refinement for LLM-Based Vulnerability Analysis
Newer and Bigger, but Safer? A Longitudinal Study of the Functionality-Security Gap in LLM-Generated Code
SCSM: A Traffic-Native Foundation Model for Transferable Website Fingerprinting
Does AI Help Cyber Attackers or Defenders? Evidence from Nonpublic Vulnerabilities and Subsequent Attacks
An Evaluation of the Semantic Understanding Capabilities of Large Language Models for Web Attack Payloads
Correct Verdicts, Flawed Reasoning: Structured Auditing of LLM-based Vulnerability Reasoning
AutoDP-LLM: Automating Data Pre-processing for Intrusion Detection Systems using Large Language Models
The AI Exposure Brief
A five-minute weekly briefing for CISOs, CIOs, AI governance leaders, and enterprise security teams.
Verified developments. Emerging agent risks. Regulatory changes. Practical control questions.
No spam. Unsubscribe anytime.