Topic
access control
Policy-aware Vector Search: A Vision for Fine Grained Access Control in Vector Databases
A new paper from arXiv presents a vision for policy-aware vector search, formalizing the problem of fine-grained access control (FGAC) in vector databases. The authors compare enforcement strategies, present preliminary findings, and identify open challenges for achieving secure, high-performance vector search in security-sensitive applications like RAG and AI pipelines.
Sovereign Execution Brokers: Enforcing Certificate-Bound Authority in Agentic Control Planes
A new security model called the Sovereign Execution Broker (SEB) introduces a runtime enforcement boundary that verifies certificate-bound execution contracts before allowing mutations in agentic infrastructure. By separating proposal, admission, and execution, SEB turns certified authority into a short-lived, revocable, auditable capability. The prototype was evaluated on AWS and Kubernetes.
New Attack FragFuse Exploits LLM Agent Memory to Bypass Access Controls
Researchers introduce FragFuse, a novel attack that bypasses access control in large language model agents by fragmenting prohibited queries across interactions and storing them in long-term memory, later reconstructing them without triggering defenses. The attack achieves an 86.3% average bypass success rate across multiple agent settings and exposes a critical vulnerability in memory-based AI systems.