Trending Repos

A 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU.

非官方 Kimi Slides Skill：让 AI Agent 生成可编辑 PPTD + PPTX，并附带本地浏览器编辑器 Unofficial Kimi Slides skill for AI agents — generate editable PPTD + PPTX with a local browser editor

Gemma 4 26B-A4B inference in ~2 GB of RAM on any M-series MacBook

Dotted thought-orb loading indicators for AI & agent UIs — six tuned states, two sizes, auto dark/light

An open-source guide for mastering WorkBuddy, an AI assistant, through real-world workflows.

Why: It helps developers effectively use WorkBuddy for various tasks and automate processes.

给 AI 编码 agent 用的 skill 集合，Claude Code 和 codex 都能跑。novel-characters：把小说拆成角色设定集——人物画像、卡通形象提示词、音色提示词、三视图。 | Skills for AI coding agents, working in both Claude Code and codex. novel-characters turns a novel into a character bible: profiles, cartoon-design prompts, voice prompts and turnaround sheets.

A visualization tool that replays coding-agent sessions on a 3D map of your codebase.

Attention is your scarcest resource. Chief is the local-first layer that guards it — turning every agent, alert, and feed into one honest call: interrupt, or not.

MAC (Multi-Agent CAD): A decoupled multi-agent framework for text-to-CAD generation via constrained test-time compute

《动手学 Pi》：沿 15 个真实 checkpoint 从零构建 Pi-style Agent

Repeatable agents-plus-code workflows, packaged as one skill, stamped into any repo. Deterministic Python owns the graph; coding agents are bounded nodes inside it.

Fast, context-efficient repository tools for AI agents (MCP)

Build AI agents that run 100% on-device. Sub-100ms latency on Qualcomm NPU. Zero cloud dependency.

雪踏乌云的 AI Agent Skills 集合

Orchestrate AI agents to find real vulnerabilities in code.

Source-backed architecture notes on building an agent

A 10-week, 30-minutes-a-day roadmap for LLM inference serving and optimization. vLLM, SGLang, quantization, speculative decoding, benchmarking.

让纯文本模型在 Codex 中无障碍调用内置看图工具（view_image）的方案，附为纯文本 LLM 设计的视觉工具包&skill ｜ Let text-only models call Codex's built-in view_image seamlessly, plus a vision toolkit&skill designed for text-only LLMs.

An open-source, ontology-based semantic context accelerator that enables AI agents to make more accurate, consistent, and explainable decisions.

A skill to kill from-scratch coding — Claude checks real arXiv prior art before it designs a new architecture.

Agent-led job application workflow with a local progress-tracking dashboard — works with any AI coding agent that can read files and follow instructions.AI Agent 驱动的求职投递工作流,配本地进度追踪看板——适配任何能读文件、听懂指令的编程 Agent。

Waku Waku! Waku agent is your personal AI agent, on your own laptop, in code you can read in an afternoon — harness + loop + memory + eval

What if your phone had no apps

Design rules to stop AI coding agents from generating generic "AI slop" UI.

A 股行情分析与AI智能投研智能体——easy stock

把照片变成超现实波普拼贴的 AI agent skill：黑白现实锚点 + 平涂色形 + 全图只有一个不可能的巨物。An agent skill that turns photos into surreal pop collages.

An OKF-backed Model Context Protocol (MCP) server delivering persistent long-term memory and SQLite FTS5 search for AI agents.

Edge-cloud Multi-Agent execution framework for embodied intelligence, connecting AI agents to real devices with verifiable actuation and entropy-bounded recovery.

梁文锋四小时投资人会议实录，因原文在微信上无法查看，特传一份github

机构级个股投研报告 Agent Skill（九章结构 + 多方法估值交叉验证），适用于 Claude Code / Cowork / Codex 等 AI Agent 工具

Soku CLI — call Soku ads/GA4/PostHog data capabilities from any AI agent or shell.

The agent skills I actually use to build software with coding agents. The PIV loop, planning, worktrees, and the meta-skills for building your own AI Layer.

To know what models don't say out loud.

Boot your PC straight into an LLM. Rust, UEFI-resident, no operating system underneath.

RL-training an AI agent to RL-train AI agents.

pgbot - Postgres intelligence for AI agents & apps

Graph-Orchestrated Agent Loop — a production-grade framework on LangGraph. Combine workflow graphs and agent loops, transpile Dify DSL to runnable code, swap wire protocols (Dify/OpenAI).

understory — memory that grows. Self-wiring, plain-markdown memory for AI agents (MCP + local models + a living graph).

Transformer architecture explained step by step - the full architecture, every attention variant, positional embeddings, and every layer inside a Transformer.

Open-source local AI workspace — advancing on-device inference.

面向 AI Agent 的可溯源 A 股板块研究 Skill，贯通政策、权威媒体、市场资金、公司与 ETF，支持实时查询、研究编排和二次应用开发。

A workbench for language model research

Speedrunning LoRA fine-tuning: frozen task, frozen hardware, public wall-clock leaderboard. modded-nanogpt for fine-tuning.

尚硅谷 AI 课程笔记

The deterministic tax engine for AI agents. Highest score ever recorded on TaxCalcBench (96% exact returns)

Evidence-aware on-prem LLM inference sizing and TCO calculator

Loom for AWS is an enterprise-grade platform for building, deploying, and operating AI agents on Amazon Bedrock AgentCore Runtime and AWS Strands Agents.

Slides Skill：让 AI Agent 生成可编辑 PPTD + PPTX，并附带本地浏览器编辑器 Slides skill for AI agents — generate editable PPTD + PPTX with a local browser editor

openfinclaw cli | ai openclaw mcp cli which works in claude code, cursor, and 20+ ai agents via mcp

Desktop shell for the pi coding agent

Privacy-first AI meeting notes & writing assistant for Mac — on-device transcription, smart minutes, action items & 60+ writing modes, running entirely on Apple Silicon

Kimi K3 free desktop - Moonshot AI K3 vs Claude Fable 5 and GPT 5.6 Sol / opus 5.0 comparison built in. kimi k3 vs fable 5 pricing and cost shown in-app. kimi k3 vs gpt 5.6 sol benchmark results reddit. kimi k3 free api key, free trial included. kimi k3 coding plan and coding agent. Download Kimi K3 desktop free today.

High-performance OpenAI-compatible gateway and token pool bridge for AI coding assistants with session lifecycle and TLS stealth

给 AI Coding Agent（Claude Code / Cursor / Codex / Gemini CLI）用的 **技能版 npm** —— 本地优先、模型无关、MCP 原生。

The NestJS-native runtime for governed AI agents

本地优先的类飞书 Markdown 编辑器,内置可审改的 AI 助手 | Local-first WYSIWYG Markdown editor with a reviewable AI agent

Creating an AI agent with LISP - the 2026 version of what made LISP "the language for AI" in the first place

Versioned operating principles for coding agents

smol is a smol agent

The world's #1 vibe coding benchmark — models go head-to-head on real engineering tasks, judged blind, ranked by Elo.

Evidence-grounded veterinary research workbench with local RAG, bounded LLM agents, auditable tool use, citations, abstention, and human review.

发一张图给你的 Codex，它自己给自己换肤 | Send one image to your Codex and it reskins itself. Agent-native skin engine for the Codex desktop app.

Auditable AI agent for electromechanical maintenance with LangGraph, RAG, fault diagnosis and human-in-the-loop review.

把 B 站视频变成学习资源的浏览器扩展（Chrome / Edge）：字幕阅读、双语对照、AI 概览、划词解释和带时间戳的笔记

本项目是《AI Agents in Action》第二版的中文译本，是一本智能体系统构建与应用的实战指南——目标不止于打造自主实体，更在于开发能够切实应对并解决现实问题的智能体。 全书从大语言模型（LLM）的基础应用出发，逐步深入到智能助手、多智能体系统与自主智能体的构建，系统讲解智能体系统的关键组成部分：角色、工具与行动、知识与记忆、检索与增强、推理与规划、评估与反馈。第二版将五个功能层作为组织原则，并新增 MCP 协议、测试驱动的智能体开发（TDAD）、智能体循环、认知与元认知等前沿主题，帮助开发者构建生产级智能体系统。

用 AI Agent 自动游玩《Turmoil》（石油大亨）并不断积累经验进化的实验项目.

面向 AI Agent 的集中式 SSH 网关，提供 MCP 接入、凭据托管、细粒度授权与全程审计。

Explorations into some of the approaches advocated by Yann LeCun, and just a more wholistic architecture (JEPA) in general

Text-To-VRMA — VRM特化型Text-To-Motionツール

Gemini 3.5 Pro Flash Free API keys via Google AI Studio. Compare 3.5 Flash vs 3.1 Pro. Run fast multimodal reasoning, autonomous coding agents, and 2M token context window. Download

Enterprise-grade Next-Generation AI Agent Infrastructure

Git-native context handoff for humans and AI agents

Research artifacts from Hyra (/ˈhaɪ.rɑː/)

Kimi K3 Code Free Desktop AI - Moonshot AI 2.8T params, 1M context. Terminal coding agent, multi-file upload, autonomous tasks. Free GitHub Copilot alternative. Download 2026.

Meta Muse Spark 1.2 Free - Terminal coding agent, 1M context,Free API, 82.9% Terminal-Bench. Repo-scale execution, parallel agents, worktree isolation. Free AI coding assistant 2026.

Production AI Systems, Agentic AI & Distributed Infrastructure

让任何 AI agent 变身个人 IP 打造策略师的开源 skill · 耳总个人实战经验沉淀 · 诊契行盘闭环 · 不碰剪辑

CoordClaw: a multi-agent system that runs AI like a one-person company. Built on management & organizational theory — teams defined in natural language, collaboration driven by a message loop (not hard-coded DAGs). Auditable work logs, human god's-eye oversight.

ADE — an agentic development environment for macOS. A roster of persistent coding agents, each with its own repo, runtime CLI, and memory.

Pithagoras — hosted web UI for the pi coding agent. Give it a task, close the browser, come back when it's done.

A practical coding agent for terminal workflows.

AI development mentor as pure prompts — make your AI coding assistant write safer, more maintainable code (supports multi-tool compatible: MIMO / Claude Code / Codex / Cursor)

Memory engine for AI agents with token-efficient, explainable retrieval beyond vector similarity.

Strata Studio lets your AI agent control your DAW so you can focus purely on making music. Instead of spending 30 minutes manually levelling tracks for streaming, removing background noise from recordings, or routing effect plugins across every track, just tell your AI agent (Claude Code, Cursor, Codex, Hermes, or Gemini) what you want.

A transferable, self-enforcing software-engineering methodology for AI coding agents — playbook, skills, contracted subagents, lifecycle hooks, and evals.

Devnors Data - Python SDK: 接口,数据API,裁判文书,Legal Case,法律法规,法条,Law Article,企业工商,Company Registry,企业年报,Annual Report,税号开票,Tax Invoice,失信核查,Dishonest Judgment Debtor,被执行人,Enforcement Debtor,关键词指数,Keyword Index,微信指数,WeChat Index,热搜榜,Hot Rank,微博热搜,抖音热搜,快递查询,Express Tracking,快递公司编号,统一查询,/v1/data/query,API Key,MCP,AI Agent

Spring AI Agent 学习

`xxg-portrait-rebuild-light` is an image-edit skill for existing portrait photos. It reconstructs the key light, fill, shadows, and background atmosphere with a director-led lighting workflow while restoring clean, healthy, low-contrast photographic skin microtexture.

A finite ontology of 115 recurring software elements and a composition framework for LLM-based coding agents

Orchestrating Mathematical Reasoning Agents with Fact-Graph Memory

Homekit gives any AI agent direct, physical control over Apple Home. Flip lights. Lock doors. Read sensors. Through a single open interface.

An agent that edits your Word and Excel files — then looks at them, through real Microsoft Office, to check its own work

A living atlas of AI companies, products, founders, investors, traffic signals, and research notes.

A free, open-source coding agent inside VS Code.

Approving is an open-source, self-hostable platform for turning coding agents into visual, reviewable, and recoverable delivery workflows. Agents run in real Docker sandboxes, exchange structured artifacts, and pause for human Approve at critical nodes.

Claude sonnet 5 free access configurations via API reverse proxies and open-source web UI wrappers. Deploy this advanced Anthropic reasoning model for complex coding, agentic workflows, and deep context analysis. Customize system prompts, bypass limits, and download

Navigational memory for coding agents: chart discoveries as you explore, cache-bust when files change

A companion for coding agents to help you understand large outputs

Built on **Agentic RAG** (Agent-driven Retrieval-Augmented Generation) technology, it not only accurately answers pre-sales and after-sales questions but also generates personalized usage reports and optimization suggestions by deeply analyzing device data.

Gradient descent, but the parameters are agents — a parallel, asynchronous framework for self-evolving agents (skills, prompts, harnesses). Diffs are the gradients; the aggregator is the optimizer.

Real-Time Multimodal AI Agent for Desktop Interaction

Technical books and research dossiers about the infrastructure that gives AI agents private workspaces, controlled execution, durable state, and safe publication.

🔥 Turn any URL into clean, structured, LLM-ready content. The open-source web fetching & extraction API.

Cairn v0.6.39 is a cross-platform desktop Git client combining repository administration, terminal-based work, and AI coding-agent workflows in one application for developers managing Git projects.

AI agent for end-to-end software development, from requirements to operations.

A tiny Rust gateway for running coding agents across model providers safely.

Custom proactive skills for Azure SRE Agent — governance, cost intelligence, architecture quality, AI workload posture

Claude Fable 5 free desktop Anthropic's 1M token model back online after the July 7 outage. Full Claude Code agent workflow, autonomous coding and deep research on Windows, macOS and Linux. No subscription required pricing info available in-app. Download today.ai-desktop

Semantic diff review for coding agents

Web UI Coding Agent | ReAct Loop | Mobile Supported

一个面向 Coding Agents 的轻量插件：帮你说清问题、冷静判断开发必要性，并认真寻找已经存在的产品、开源项目和可复用组件。

Local HTML-artifact review tool for coding agents — render a plan as HTML, annotate it in the browser, send feedback back to the agent. Installable as an agent skill.

DocVectra is an enterprise-grade intelligent knowledge base system built on RAG (Retrieval-Augmented Generation) technology, providing precise knowledge retrieval and intelligent Q&A services for vertical domains.

A portable AI-assisted Knowledge OS starter kit for Obsidian, Codex, Claude Code, and compatible AI coding agents.

Take the 100 Days of AI challenge - learn one AI concept a day, from machine learning basics to AI agents in production, in 100 days.

Context Diet is a Claude Code agent skill for measuring system-prompt overhead, reviewing tool payload size, and reducing unused definitions through controlled settings.json configuration.

The CUSTODY(TM) Framework for AI Agent Containment

DaVinci AI v2.0 is a local web application for finding, analyzing, and researching personal notes and documents with retrieval-augmented generation in a focused 2026 knowledge assistant platform.

Unified Agentic AI and Business Intelligence Platform

Interactive 3D visualization platform for exploring transformer architectures, tensors, and real-time LLM inference.

专为 AI 玩家设计的单文件卡牌肉鸽 🌙

MCP server for Unreal Engine 5.6/5.8 — token-efficient Blueprint reading, editing, and a persistent project index for AI coding agents

macOS & Windows Codex Desktop 梦幻换肤 Skill｜由世事宜AI免费提供

Agent hands for your Mac — AX-first desktop control CLI for Grok Build and coding agents

从Claude Code、Codex、Reasonix等Agent工具导入迁移历史消息，并在DeepSeek Harness(DSH)中继续对话

python notebook facial recognitition

Direct Blender with an AI agent from your terminal.

hummingbird is a lightweight, zero dependency runtime for massive open source Mixture of Experts (MoE) language models. It unifies SSD, RAM, and VRAM into a single intelligent memory hierarchy, enabling inference of models like GPT-OSS 120B, GLM, DeepSeek, Qwen, and more on consumer hardware

本地 Grok Runtime，为多种 AI 编程 Agent 提供可管理的交互式会话 | Local Grok runtime for manageable interactivesessions across AI coding agents

ash4d.com is a self-hosted private AI platform for k3s Kubernetes, combining local Qwen3 inference, retrieval-augmented generation, agent tooling, and GPU-enabled services in a GitOps-managed environment.

Tells parallel AI coding agents when their in-flight work would break each other — before either lands. Catches the case git merges cleanly.

AI-powered Labour Law Violation Detection System using Weak Supervision, BiLSTM, Graph Neural Networks (GNN), and Explainable AI to identify labour law violations.

Antares — a self-hosted AI agent: Go backend, React dashboard, terminal UI, and messaging gateways.

An open experiment in memory for AI agents — the whole memory, always in the prompt.

Run GLM-4.5-Air (110B) on a 16GB-RAM consumer machine - identify the best memory allocation to overcome standard hardware limitations in Local LLM applications. Placement beats budget. Falsification-Tested laws, probes and recipes for LLMs on commodity hardware

一个生成 「影视飓风(Mediastorm)」风格中文口播稿 的 Claude Code / Agent Skill。

Official Repo for "AgenticASR: Refining Speech Recognition in Real-World Scenarios via an Agentic Approach"

An autonomous AI agent written in Go, designed to run entirely on the 32MB RAM of an obsolete OpenWrt router. It acts as an always-on smart home orchestrator with Wake-on-LAN, Telegram, and Google Workspace integration.

Part of the KDAN ecosystem, DocSlight offers document parsing, OCR, and data extraction that turn PDFs, scans, images, and Office files into structured outputs for RAG pipelines, AI agents, and enterprise document automation.

macOS coding-agent control center: one objective, routed to direct execution, a bounded loop, a swarm, or a dynamic DAG. Local only, no external writes.

Your own cloud coding agent. Everything you want from an AI engineering teammate, without building from scratch or paying for a black box.

AI Security Monitor — Real-time threat detection, prompt injection defense, and behavioral analysis for LLM-powered systems

Implementation of RoboTTT proposed by Yunfan Jiang et al. of Stanford and Nvidia

Gogh is a source-cited Obsidian operating brain that turns six frontend design skills into one advisable stack for AI coding agents.

Local web console for Pi Coding Agent — models, agents, sessions, usage, OpenVL

梨园 Liyuan — 以 AI Agent 为主体的角色扮演应用：记忆账本、决策卡、自建面板、世界线，兼容角色卡/世界书

一个基于 .NET 10 构建的现代化 AI Agent CLI，采用 Runtime First、Plugin First 架构，支持多模型、MCP、Skill、Tool、流式终端交互和可扩展插件系统。

The Kubernetes control room—built in Rust, ready for engineers and AI agents.

Claude Code Sonnet 5 desktop free AI coding assistant on Windows, macOS and Linux, no API key needed. Benchmarks above Sonnet 4.6 at lower cost than Opus 5. Beats GitHub Copilot on context window, compares favorably vs Fable 5 on speed. Download and start coding.

A Codex skill that keeps human outcomes and creative intent ahead of overengineering.

A safe, auditable relay between planning AIs and coding agents—without repetitive copy-pasting.

A concise, opinionated set of design rules for coding agents to produce better user interfaces.

A hypervisor and sandbox cloud for self-hosted AI agents and RL

An agentic LLM-powered knowledge assistant that enhances RAG capabilities through automated entity extraction, structured data analysis, and SQL-based reasoning to deliver accurate answers for complex, multi-step, and analytical queries.

python notebook anti spoofing facial recognition

Reusable AI agent skill for cloning, mirroring, offline-validating, and rewriting complex web experiences.

AI agent autonomously optimizing RAG retrieval pipelines against your own documents

Agent-driven SAP/ABAP knowledge base engine that turns S/4HANA custom objects into citable Markdown/Obsidian context for humans and AI agents.

A disciplined, delegation-driven operating system for AI coding agents. npx contractor-kit.

A visual benchmark testing whether leading coding agents repeat the same design patterns across 100 neutral website briefs.

Offline, local-only context migration skills across 9 coding-agent sources × 9 destination hosts (registry-derived 81 cells). Stdlib-only; inert handoff for fresh sessions, not live restore.

Lydia is a local AI coding agent for the terminal — reads and edits your code, runs commands, and drives git through a local Ollama model. No API keys, no cloud, nothing leaves your machine.

An open-source AI agent for the browser. Bring any model. It plans in the side panel, acts on real pages through tools, and verifies every claimed result against the live DOM — with guardrails and reusable skills.

A terminal interface for Nanocodex

International secret & PII redaction for JS/TS: logs, prompts, HTTP, datasets. 20+ languages, checksum-validated national IDs, reversible LLM layer. Zero deps, ReDoS-safe.

Workflow engine, JSON control-flow tool, and live terminal viewer for the pi coding agent

Anvil — agent-native multi-genre game engine for AI coding agents

Human-gated, capability-slotted collaboration workflow (SOP) for coding agents — discovered rosters, condition-based roles, cross-vendor review, signoff gates

Mindlas catches your coding agent drifting before the bad code lands. Real-time tracking and correction for context rot, unverified done claims, patch sprawl, and tool loops in your Claude Code session.

Repo Steward AI: safety-conscious PR review, issue triage, release notes, and verifiable maintainer pilots with the OpenAI Responses API.

Find out why your Codex hooks aren't firing

The open-source experience learning layer for AI agents.

The video editor you own, not rent — edit clips from your coding agent (Claude Code, Codex, Gemini); e.g. burn in stylized subtitles.

OpenABCode is an LLM-routing coding agent that dynamically routes tasks to the best-suited models

SpaceXAI's coding agent harness and TUI. Fullscreen, mouse interactive, extensible. WITH TELEMENTRY REMOVED AND INTERNAL JAILBREAK ADDED!

Implementation of Light Loco Parkour, Versatile Perceptive Whole-Body Locomotion via Multi-Skill Distillation, out of Light Origins Labs

What AI runs here, and what it costs. A local scan of your AI tools, sites, tokens and spend.

Multi-session IDALib MCP router for coding agents. Analyze multiple binaries in parallel with IDA-compatible reverse engineering tools.

AI Kubernetes CLI: natural language → reviewable plan → approve before apply

从零写出你自己的 coding agent：对照 pi / codex / grok-build 三个工业级开源实现

python notebook super resolution

A tree-guided development plugin for long-running coding agents.

OpenAI-compatible proxy for Z.AI (GLM models)

面向义务教育教材的 AI 知识抽取与交互式知识图谱平台，支持 PDF 解析、知识点提取、向量检索、前置关系构建及 MCP 查询服务。

pytest for LLM apps: record API calls once, replay them forever, and test meaning without flaky live runs.

An agentic AI tool for searching, screening, and ranking NIH funding opportunities based on research interests and proposal relevance.

Drop-in skills that stop AI coding agents from shipping garbage — kill-slop, security-review, PR hygiene

A desktop AI coding assistant that works with your local projects. Ally helps you understand code, edit files, search a workspace, manage tasks, and complete development work through conversation.

Local AI coding-agent history + API cost analytics (loopback-only).

Implementation of RoboTTT proposed by Yunfan Jiang et al. of Stanford and Nvidia

A safe local browser wizard for installing an OpenAI OAuth sidecar beside self-hosted n8n.

把本机 Codex 工作记录生成可分享的 AI 打工小票：一行命令、完全本地、隐私优先、中英双语、主题切换，并支持扫码导入微信小程序。

100 道大模型面试题与 100 张 SVG 图解，覆盖 LLM、RAG、Agent、微调、推理优化和评估。

python notebook ReId project

完全开源的一站式 AI Web 平台（99AI 插件版）：多模型对话、Agent 工作流、MCP 工具、用户额度与支付运营。NestJS + Vue3 全栈完整源码，Apache-2.0，支持 Docker 私有化部署。

Compact USB-C development board concept built around the Microchip ATSAMD11D14A, developed by UI.