A new research framework called MOSAIC successfully compromised AI coding agents in 96.59% of 2,525 attempts by exploiting CLI command-composition risk (CCR), a novel attack category that does not rel…
University of Illinois Urbana-Champaign researchers have released an open-source AI teaching assistant on HuggingFace that runs 11 separate models in parallel for text/image retrieval, generation, mod…
A 2026 study from UC San Diego, UC Berkeley, Texas A&M, and UIUC found that context pruning—masking stale observations from an AI agent's history—improves accuracy by up to 11.7 points in some setups …
NVIDIA and academic researchers introduced ASPIRE, a self-improving robotics framework that achieves 31% zero-shot success on LIBERO-Pro long tasks. The system uses a coordinator-actor architecture wi…
Researchers introduced TimeClaw, a framework that equips generalist AI agents with tools for analyzing time series data within real-world contexts like energy, finance, and weather. The system integra…