Full-text search across 17142 articles. Combine with topic and date filters; results sorted by relevance.
A developer built the official Postman data-source connector for Cognee, an open-source memory layer for AI agents, during Mergetober 2026. The connector ingests Postman collections into Cognee as queryable knowledge gra…
A paper submitted to arXiv on 2 October 2026 reports that an observer at the split in split learning can rebuild most of a client's text from the traffic, with gradients adding measurable leakage: on GPT-2, an attacker h…
Yaala Labs released Agent Kernel, an open-source, framework-agnostic runtime for developing, testing, and deploying AI agents across frameworks including LangGraph, OpenAI Agents, CrewAI, and Google ADK. The runtime offe…
A 2 October 2026 arXiv paper evaluating fifteen prompt-injection detectors, including Meta's Prompt Guard 2, found that public benchmark scores transfer poorly to real LLM agent deployments: the best detector on the BIPI…
A paired-replay testbed measuring task-scoped authorization in tool-using LLM agents found that broad-bearer credentials permitted harmful tool execution in 8.9% to 37.8% of valid attacked post-exposure decisions across …
Researchers introduced Queen, a 4B-parameter chess-language model that plays at Grandmaster level and explains its moves, gaining over 900 Elo points (1782 to 2697) across seven iterations of training, according to an ar…
Researchers Shashank Kirtania and co-authors introduced BREW (Bootstrapping expeRientially-learned Environmental knoWledge), a framework that distills an LLM agent's past interaction trajectories into a structured, retri…
A September 17, 2026 arXiv paper proposes the Output-Space Hypothesis, an enumerative equivalence-checking approach implemented in a system called Dirigo that flips the standard quantifier order to check a single output …
Meta released the open-source Muse Gadgets SDK, which connects ESP32 microcontrollers and Raspberry Pi or Linux machines to its Muse AI agent over Bluetooth LE, with the GitHub repository reaching 400 stars within 72 hou…
A 2 October 2026 arXiv paper introduces STEER-Bench, a 101-task benchmark across 9 domains, and reports that branch steering attacks succeed against 94.4% of standard Computer Use Agents and 89.5% of vanilla Dual-LLM CUA…
A hybrid digital-analog deepfake video detection framework combining a lightweight digital front-end with a spatially multiplexed optical decoding back-end achieved 97.79% average detection accuracy, 99.86% sensitivity a…
A September 30, 2026 arXiv report by Ling Min Serena Khoo systematically compares several nearest-neighbour retrieval variants for molecular fingerprint prediction from MS/MS spectra, showing how differing assumptions ab…
A July 15, 2026 arXiv paper titled "Harness Engineering: Anatomy, Architecture, and Evolution of Coding Agents" audited the source code of eleven production coding harnesses — Claude Code, Codex CLI, Gemini CLI, Mistral …
Anthropic shipped Claude Code Mods on October 1, a TypeScript extensibility layer that runs inside the agent's execution pipeline and can intercept tool calls, rewrite the system prompt, redraw the terminal UI, and repla…
Google Research published TurboQuant on March 24, 2026, a KV-cache compression algorithm that shrinks long-context memory to 3 bits per value with no accuracy loss, cutting a 16GB cache for Llama-3.1-8B at 128K context t…
A difference-in-differences study of Google's AI Overviews (AIO) found that default AIO availability reduced English Wikipedia's external search traffic by 5.45% versus German and 4.82% versus French, according to the ar…
A June 8, 2026 arXiv paper (2606.08919) argues that human-in-the-loop approval gates for LLM agents rest on two false assumptions — that a ground-truth notion of "risky" exists and that the human reviewer is a perfect, i…
SuperLocalMemory 4.0, a governed local-first memory operating system for AI agents submitted to arXiv on 8 Aug 2026 by Varun Pratap Bhardwaj, reports that 10 of its mechanisms were implemented and reachable on a live cal…
Researchers posted a paper on arXiv on 1 October 2026 introducing LoopCD, a training-free contrastive decoding framework for looped Transformers that contrasts the final prediction with an earlier recurrent pass. LoopCD-…
Open source inference engine Strata, built by a developer known as Niko, runs Alibaba's 125 billion parameter Qwen3.8-Flash-Next mixture of experts model on a single consumer GPU with as little as 8GB of VRAM, shipping v…