2026年11月10日 技术热点总结

项目简介：Agent skills

核心介绍：A collection of skills for AI agents and chatbots. Each skill is a directory containing a SKILL.md file with instructions for the agent. topic of your choice using the Socratic method: a diagnostic, a skill map with levels 0-5, one question at a time, spaced review, and a final assessment. With a coding agent that supports the common SKILL.md format, copy the skill directory into your agent’s skills folder, then ask the agent to use it

项目数据：⭐ 2,053 Stars，🍴 112 Forks

项目简介：Run your own Switch games natively on iPhone, iPad and Mac: game code rewritten to ARM64, graphics direct to Metal, one app for the whole library. Bring your own games and keys. 把你自己的 Switch 游戏原生带到 iPhone / iPad / Mac。

技术栈：C#、apple-silicon、arm64、game-porting、ios、ipados、macos、metal、s2arm

核心介绍：Platforms: iPhone / iPad (iOS / iPadOS 18 and later) · Mac (Apple silicon, macOS 15 and later) · Build on Mac or Windows · Android / HarmonyOS coming soon · License: Apache-2.0 · Bring your own games 简体中文 · Download the launcher · All releases · Evidence

项目数据：⭐ 1,742 Stars，🍴 …

项目简介：An MobileGestalt editor for iOS 27.x, featuring region-code tweaks, device presets, backups, and support for 7 languages.

技术栈：Swift

核心介绍：GestaltX English · 简体中文 A MobileGestalt editor for iPhone and iPad. Developed by huamidev.

项目数据：⭐ 1,508 Stars，🍴 571 Forks

项目简介：A GPU task manager for Windows, Linux, and WSL: per-process VRAM, GPU activity, and LLM model/KV-cache insights (Windows AMD and NVIDIA).

技术栈：Python

核心介绍：Windows + NVIDIA/AMD, with Linux and WSL support (NVIDIA everywhere, AMD and Intel on native Linux; see Linux and WSL) and macOS on Apple Silicon (see macOS). A terminal UI that connects per-process GPU memory and engine activity with loaded runtimes, model files, and LLM buffer breakdowns. Task Manager says “15 GB used.” This tries to answer: *by what? Model weights? KV cache? Your browser? The desktop compositor? And w…

项目简介：Monitors your RAM and reports any “quite terrible” unused capacity directly to the CEO of OpenAI.

技术栈：Shell、bash、cron、dram、email、humor、linux、meminfo、memory

核心介绍：Read the sample emails at ram-altman.lol, no install needed. There is RAM in your computer. Right now. Just sitting there, in a *personal* computer, being personal. You may have thought of this as “a computer working normally”. On a recent episode of Cheeky Pint, the CEO of OpenAI offered another view: > “From a global allocation perspective, actually, it’s quite terrible to have everybody have a bunch of the …

项目简介：Let AI agents use your phone, with you in the loop.

技术栈：Go

核心介绍：Pinky Let any AI agent use your iPhone or Android phone, from macOS, Windows or Linux, with you in the loop. Download · Get started · FAQ · Discord English · 简体中文 · 日本語 A real Claude Code run on an Android 16 emulator. Pinky asks before payments by default; this run also asks before messages.

项目数据：⭐ 1,424 Stars，🍴 152 Forks

Automating printed circuit board (PCB) routing requires coordinating board-level connectivity decisions with geometric execution under strict design rules. Although recent agents can interact with native routing tools, reliable routing on complex boards remains challenging. We present PCBRouteAgent, a multimodal PCB routing agent inspired by how human engineers coordinate global strategy with local execution. A Global Routing Planner combines board images and structured electrical and geometric information to specify connection topology, routing order, corridors, and layer transitions. A De…

Large language models reinvent the wheel all the time. They conflate task-solving with the execution of the solutions, which forces them to solve every problem from scratch. We show that once a task has been reasoned through, the application of its solution to a new problem instance can be done very efficiently and with little involvement from language models. We introduce agentic reasoning caching, a method for constructing robust CPU-executable agentic reasoning caches (ARCs) from LLM reasoning traces. ARCs show on-par performance with frontier LLMs across a range of difficult mathematica…

Intra-weight injection lets one model change another model's behaviour by editing its weights directly, without gradients. Such edits can range from altering lexical associations to enforcing output formats or updating factual knowledge. Despite growing interest in these approaches, there has been no systematic way to evaluate how well models perform them. We introduce IWIBench, which measures a large editor model's ability to modify the weights of a smaller target model (typically Gemma 3 4B IT) to achieve one of ten classical text-editing objectives while preserving its other capabilities…

Financial document recognition is often assessed using character and word error rates, although these metrics do not express whether a system preserves the financial facts on which people and organisations act. This paper develops a human-centred, risk-constrained framework for evaluating financial document recognition as evidence recovery and decision support. It defines typed financial facts linked to visible source regions and separates verbatim transcription, deterministic normalization, and business interpretation. A layered quality model covers text, localization, layout, tables, fiel…

Language model deployment often requires different trade-offs between competing objectives as user preferences change, yet standard post-training methods typically produce a single policy for a fixed objective weighting. We introduce LambdaMix, a new parameter efficient method for learning a preference conditioned family of policies with group-relative online reinforcement learning. LambdaMix jointly trains one low-rank adapter per objective over a frozen base model. One preference vector selects the adapter mixture that generates responses and weights the separately normalized rewards used…

Autoregressive decoding relies on the KV cache, whose memory footprint grows linearly with context length and caps batch size and throughput. Existing post-training methods exploit redundancy across tokens, layers, and feature dimensions, or lower numerical precision. We identify a complementary axis: within each attention head, values are largely predictable from their corresponding keys. We exploit this with value cache amortisation, which fits a lightweight per-head predictor at test time and stores residuals only for the tokens it reconstructs worst, giving 4times near-lossless value co…