月报
2026-11 月报精选：Anthropic takes steps to blo；kajisho5/ffmpeg-skill；ATLAS-Bench: An Agentic Tool
Local video editor for coding agents: 42 FFmpeg tools with a machine-readable contract and an MCP server, for Claude Code, Cursor and Codex. Cut, captions, reframe, loudness, platform exports. No cloud, no API keys.
414 art-style prompts as a ComfyUI node with live preview + a split-slider gallery to SEE the whole library on one subject 💅
arXiv:2610.39438v1 Announce Type: new Abstract: Reliable local search is essential for LLM agents to support everyday decisions and longer plans by finding places that meet users' needs at given times and locations. Yet existing benchmarks provide limited coverage of whether sear
arXiv:2610.39377v1 Announce Type: new Abstract: LLM agents increasingly carry out multi-turn tasks across everyday applications, yet users often see only the agent's report rather than the observations, API calls, and state changes behind it. Existing deception benchmarks have st
MCP server for Home Assistant — 323 tools across 29 domains: entities, automations, dashboards (sections-aware), Card Builder, energy, ESPHome/LVGL, HACS, Supervisor, git config versioning, BM25 tool search, aggregated s
Building an autonomous agent requires a model, a harness, and domain-specific tools. To deploy an agent more securely, developers need a runtime that enforces policies, along with optimization and observability tools for reliable execution. For production agents, the challenge li
arXiv:2610.37684v1 Announce Type: cross Abstract: Large Language Models (LLMs) have rapidly become important tools for science, education, medicine, and professional work. Despite their apparent simplicity from the user's perspective, effective and responsible use requires an und
Mark Zuckerberg's Meta made AI look so stinking cute by giving it a squishy little face - and this holiday season, he wants to sell you a Muse Charm so you can carry your jolly AI helper around like a Tamagotchi. But I won't be buying one, because we have Muse Charm at home . It
📌 一句话摘要 MovieFlow 创始人梁巍将 16 年的电影工业经验封装成 AI 智能体，让普通人也能轻松创作高质量视频，推动内容生产从精英垄断走向全民参与。 📝 详细摘要 梁巍，前电影制片人，现为 AI 视频平台 MovieFlow 的创始人。他将复杂的电影制作流程拆解为多个智能体，通过多模态大模型的协同，让用户只需输入简单构想即可生成连贯的视频内容。这一创新不仅降低了创作门槛，还显著提高了生产效率，为影视行业带来了新的可能性。 💡 主要观点 多智能体协同是核心竞争力 MovieFlow 通过将剧本、分镜、摄影、灯光等环节拆解为独立的 AI 智能体
LLM observability for AI agents Discussion | Link
See how you really vibe code with Claude Code & Codex Discussion | Link
AI is expanding beyond the data center as developers build a new generation of agents to act at the edge: monitoring driver attention and cabin conditions, helping robots understand their surroundings and plan tasks, and coordinating fleets of machines on production lines. These
📌 一句话摘要 本文深入探讨了 AI 编程 Agent 的核心架构，强调了 Harness（执行框架）在决定 Agent 实际表现中的关键作用，并分析了开源社区和官方解决方案的最新进展。 📝 详细摘要 随着大模型能力的趋同，Harness 成为区分优秀与普通 Agent 的决定性因素。文章通过具体案例（如 Can Bölük 的 Hashline 实验）展示了优化编辑工具如何显著提升成功率，并介绍了 Oh-My-* 生态中各项目的创新功能。同时，分析了 Anthropic 的 Claude Managed Agents，其通过解耦 Brain、Hands
arXiv:2610.38126v1 Announce Type: cross Abstract: Web assistants increasingly answer questions about a user's private documents by pairing a small on-device model with a cloud LLM behind a Web API. In collaborative decoding, the document stays on the client, and the cloud receive
arXiv:2610.39230v1 Announce Type: new Abstract: Financial AI systems must support decision-bearing questions involving payments, collections, cash flow, and compliance while ensuring numerical accuracy, traceability, and auditability. Large language models (LLMs) enable natural l
arXiv:2610.38691v1 Announce Type: new Abstract: Verifying numerical claims in research written by AI agents is an important but challenging problem. Recent benchmarks show that many agent-written papers fail integrity checks, but such verdicts, often produced by a single LLM judg
原作から脚本・画像生成・組版・PDF出力まで進める、Codex向け漫画制作ツール
arXiv:2610.21565v3 Announce Type: replace-cross Abstract: On-device deployment of large language models (LLMs) is increasingly adopted due to latency and privacy requirements, but exposes proprietary models to reverse engineering and intellectual property theft. Trusted Execution
arXiv:2610.38326v1 Announce Type: new Abstract: Ontological knowledge underpins large language model (LLM) performance on knowledge-intensive tasks. However, existing evaluations of LLMs' ontological understanding largely rely on behavioral correctness, leaving the underlying mec
arXiv:2610.40563v1 Announce Type: new Abstract: This study investigates the capabilities of Large Language Models to simulate counselling clients in educational role-plays in comparison to human role-players. Initially, we recorded role-playing sessions, where novice counsellors
arXiv:2608.10319v3 Announce Type: replace-cross Abstract: As human developers increasingly collaborate with large language model (LLM)-powered coding agents, adapting to individual developer preferences, potentially via agent skills, may offer way to reduce repeated guidance and
arXiv:2609.04127v2 Announce Type: replace Abstract: Large language models are increasingly used to support organizational decisions. Despite their rapid and widespread adoption, users often lack a principled basis for assessing whether to rely on a specific recommendation. We int
Custom nodes pack for ComfyUI This custom node helps to conveniently enhance images through Detector, Detailer, Upscaler, Pipe, and more.
arXiv:2610.40023v1 Announce Type: cross Abstract: Automatic operator design (AOD) has increasingly adopted large language models (LLMs) to generate problem-specific search and variation operators for multi-objective combinatorial optimization (MOCO). Yet most methods evolve the o
OpenAI will honor a shared robots.txt signal to help website operators communicate their AI training preferences to participating providers.
arXiv:2610.38082v1 Announce Type: cross Abstract: Feature extraction from unstructured data typically relies on opaque embeddings unsuited to interpretable decision-making. Interpretable features, on the other hand, typically require costly, and often infeasible, hand-engineering
arXiv:2610.39275v1 Announce Type: new Abstract: Engineering design frequently involves redesigning a prior solution to adapt to changes in context, such as material prices or system requirements. Redesign often leverages simulators too complex and/or computationally intensive for
arXiv:2610.38392v1 Announce Type: new Abstract: Self-improving LLM agents edit their own harness, the prompts and rules around a fixed model, and keep an edit only if an acceptance gate scores it higher on a batch of tasks. Such a gate sees what the harness does under the current
arXiv:2608.24069v2 Announce Type: replace Abstract: LLM-based multi-agent trading systems are moving from research prototypes to live deployments that control real assets. They ingest news and social media from the open web and pass it between agents, so a single corrupted signal
arXiv:2610.38391v1 Announce Type: new Abstract: Outcome-based reinforcement learning provides coarse supervision for long-horizon LLM agents: failed trajectories may contain correct steps, while successful trajectories can contain errors. Trajectory-level skills can also shift te
arXiv:2610.40370v1 Announce Type: cross Abstract: Large language model (LLM) agents for personal health read wearable data and recommend, but do not learn about the individual in a way that can be audited or falsified. We present EpiCoach, an agentic system that treats one athlet
arXiv:2610.39621v1 Announce Type: new Abstract: The integration of Large Language Models (LLMs) into agentic architectures has substantially increased token consumption, raising concerns about their energy footprint. Yet, estimating the energy consumption of proprietary API-based
arXiv:2609.01202v2 Announce Type: replace-cross Abstract: Clinical trials are essential for evidence-based medicine and drug development, but many fail because of insufficient patient enrollment. Although there is growing interest in using AI to support patient recruitment, exist
arXiv:2610.38917v1 Announce Type: cross Abstract: Large language models (LLMs) have rapidly become a default source of help for students in science and engineering courses. In foundational engineering mechanics courses such as statics, students routinely use these tools to check
arXiv:2610.39417v1 Announce Type: cross Abstract: Large language models (LLMs) have shown strong performance on software engineering (SE) tasks, driving the need for high-quality datasets for reliable training and evaluation. SWE-bench-style datasets (SWE datasets) are widely use
arXiv:2610.19345v2 Announce Type: replace Abstract: AI agents use a large language model (LLM) to plan and act over many steps. Which model should an agent use, so that its tasks are completed reliably with the least energy, cost and time? We treat this as a multi-criteria decisi
arXiv:2609.17989v3 Announce Type: replace-cross Abstract: Consumers increasingly delegate product search to AI shopping agents powered either by generative large language models (LLMs) that deliberate in natural language or by non-generative decision models that output calibrated
OpenAI is strengthening violence prevention in ChatGPT with expanded safeguards, earlier interventions, and a new commitment to public safety reporting.
Use Claude, Anthropic's AI assistant, right from your terminal. Claude can understand your codebase, edit files, run terminal commands, and handle entire workflows for you.
It’s a quiet day, so as we have done once per conference, we’ll do a behind the scenes look at AI Engineer , now serving 6M views a month for 1.8M monthly unique viewers . CODE is the last flagship AIE of the year (there are also 2 partner conferences coming up - AIEi Shanghai ta
I bought a no-name "DQ10" Android TV stick (Allwinner H618) and asked Claude Code to "reflash it with clean firmware and make it safe". About two days later: the stick runs mainline U-Boot + Armbian (kernel 6.18.53) from its own eMMC; the factory Android is completely wiped, and
arXiv:2609.19906v2 Announce Type: replace-cross Abstract: Closed-loop robot policies require observation processing, state management, and situation-dependent branching, making them costly to design and tune manually. Although coding agents increasingly support control-code gener
arXiv:2610.38252v1 Announce Type: cross Abstract: Music audio generation spans instrumental clips, full songs with vocals, edits of existing recordings, and systems that play alongside a musician in real time. This monograph explains the methods behind these systems for readers f
arXiv:2610.40556v1 Announce Type: new Abstract: Clinical genomic analysis requires experts to integrate evolving guidelines, databases, and computational evidence. As sequencing has become increasingly accessible, this complex and labor-intensive process has become a major bottle
Run rs-matter on Espressif chips with ESP IDF
Your smart lights show when Claude Code or Codex needs you. Discussion | Link
Mosaico for CiviCRM gallery manager
See what every AI coding agent is doing, with whatever is waiting on you at the top. A Claude Code plugin that also follows Codex, Gemini CLI, Cursor, OpenCode, Grok Build, Hermes Agent and Muse Code.
50x your Claude usage with cheap-ag. Opus plans, a dirt-cheap model writes every line, and tests decide what ships. Your $20 plan stops hitting limits. One MCP server for Claude Code.
Canto Arte Mosaico é um site institucional desenvolvido para o artesão Fernando Paraboli, especializado em mosaicos artísticos. O site apresenta um design responsivo, utilizando Bootstrap 3, jQuery, Owl Carousel e AOS pa
🌃 Capri — Edgerunners-inspired AI agent for your desktop. Works with Claude Code, Codex, Antigravity & more.
Give a local AI coding agent a remote hand over SSH: the brain stays local while an SSH tentacle reaches into a remote Linux server to read/run/diagnose, behind tiered safety. Reverse Remote-SSH, packaged as a Claude Cod
Brightspace D2L MCP server for Claude, ChatGPT, Cursor and Copilot. Knows what you already handed in, so it only tells you what is actually left. Due dates, grades and course files at any school with normal SSO sign-in.
SentinelX agent: outbound-only agent that exposes your Linux, macOS, and Windows servers as MCP tools to Claude.ai and ChatGPT
Job Search MCP for finding remote jobs with Claude, Cursor, or any AI agent. Search listings, read descriptions, compare roles, and find where to apply.
Open source Grok Bot alternative, Bring your own agent and subscription: message Claude Code, Codex, Gemini CLI, Cursor, Copilot and 40+ AI agents as bots on your own computer, from iPhone, Mac, Windows, Linux or termina
Mac app and JSON CLI to monitor every Claude Code, Codex, Cursor Agent and Pi session on your machine. See which agent needs permission, which is working, and which is idle. Works with any terminal or IDE. Free, open sou
Give your coding agent subagents in your other agent CLIs and accounts, all sharing the same tools. Plugin for Claude Code, Codex and Antigravity; MCP server for any other client.
13 Claude Code skills for video production (transcribe / translate / dub / multicam / subtitles / reframe) + WeChat publishing. Compatible with Claude Code, OpenAI Codex CLI, Cursor, Gemini.
Official MCP server for Frihet ERP — 157 tools for AI-powered invoicing, tax & e-invoicing compliance (VeriFactu, TicketBAI, Facturae), banking, CRM, HR/payroll & POS. Works with Claude, ChatGPT, Cursor, Windsurf, Cline,
基于 ESP-Mosaico（ESP32-S31 + 480×480 显示屏 + 交互子板）复刻的经典弹幕生存游戏：一条命、不能射击、只用左右移动，在越来越疯狂的字弹地狱里坚持 60 秒。
app de coleta de dados parques
A collection of AI agent skills for working with Expo projects and Expo Application Services
UEG 数字生命卡 · AI Passport 固件源码 (ESP32-C3 / ESP-IDF v5.5) · AI 余额看板 + F1 指挥台
Distributed image and video generation in ComfyUI for two NVIDIA DGX Sparks.
Sonos server for my 5 year old to control his speaker using an esp32s3 M5Stack CardPuter
Локальная студия генерации кинематографичных видео со звуком на MiniMax H3 — Windows, NVIDIA GPU, полностью offline.
Open-source 3D camera, character & motion previs studio replacing Higgsfield 3D-Jutsu for AI video.
Train LoRAs inside ComfyUI: image and video models, live previews, on NVIDIA or AMD.
An open source `vercel` like deployment platform for Comfy UI
Runway ML Pro is the paid Runway plan for making video. A short text line or one still picture goes in, and the plan spends its credits to produce a clip. Pro is the higher credit plan on that account. The account stores
AI agent to iteratively create visual content via comfyui workflows
GitHub's AI coding agent that picks and combines models Discussion | Link