GitHub 每日热点项目(2026年10月18日)
-
1 anthropics/claude-plugins-official
由 Anthropic 官方管理的高质量 Claude Code 插件目录。
Official, Anthropic-managed directory of high quality Claude Code Plugins.
Python 41,763 4,465 新增 +1,397
项目简介这是一个 Anthropic 维护的 Claude Code 插件目录,分为内部插件(/plugins)与第三方/社区插件(/external_plugins)。用户可通过 Claude Code 的 /plugin install 或 Discover 浏览安装。README 提醒用户在安装前自行信任并审查插件;插件名发布后不可变,改名需通过 renames 映射迁移。仓库还说明插件基本结构及 skill-bundle 插件的声明方式。
-
2 MadsLorentzen/ai-job-search
在你自己机器上运行的求职助手。一个基于 Claude Code 的 AI 求职申请框架:评估职位、定制简历、撰写求职信、准备面试。Fork 后自行掌控。
The job search that runs on your machine. AI job application framework built on Claude Code: evaluate postings, tailor CVs, write cover letters, prep interviews. Fork it and own it.
Python 50,413 16,940 新增 +1,365
项目简介这是一个基于 Claude Code 的求职申请框架,核心命令为 /setup、/scrape、/apply,另有十个扩展命令(如 /interview),覆盖从自我画像、职位搜索、适配评估到简历/求职信生成与面试准备。其 /apply 使用起草-评审双代理流程,并用 LaTeX 编译、检查 PDF 和 ATS 可读性。README 称作者用其投递 69 份申请、获得 20 个首轮面试并签约;职位抓取技能以丹麦市场为主,但工作流可移植到其他地区。
-
3 affaan-m/ECC
面向 agent harness 的性能优化系统。为 Claude Code、Codex、Opencode、Cursor 及其他平台提供技能、本能、记忆、安全和研究优先的开发。
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
JavaScript 282,514 41,987 新增 +1,145
项目简介ECC 是一套编码 agent 的工程系统与工具箱,提供 71 个专用 agent、311 个技能和 97 个兼容命令,并包含 hooks、会话记忆、持续学习、规则和 AgentShield 安全扫描。README 称其与 Claude Code 配合最佳,有 Codex、Qoder 的原生路径,并为 Cursor、OpenCode、Gemini 等提供适配器。安装可通过 universal 引导脚本或 Claude Code 插件 ecc@ecc 完成;MIT 开源。
-
4 HKUDS/CLI-Anything
“CLI-Anything:让所有软件成为 Agent-Native” —— CLI-Hub:https://clianything.cc/
"CLI-Anything: Making ALL Software Agent-Native" -- CLI-Hub: https://clianything.cc/
Python 54,062 4,905 新增 +574
项目简介CLI-Anything 旨在通过命令行让软件可被 AI agent 调用,并提供 CLI-Hub 作为浏览、安装和管理社区 CLI 的注册表。它包含一个七阶段生成器,可从源码/GUI 分析、设计、实现到测试、文档和发布来生成 CLI harness;也支持 refine、test 等命令。README 列出的平台包括 Claude Code、Pi、OpenClaw、Cursor、Codex 等,并声明多个社区 CLI 适配如 GIMP、Blender、Zotero、Godot 等。
-
5 K-Dense-AI/scientific-agent-skills
将任何 AI agent 变成 AI 科学家。面向科学的第一 Agent Skills 库,全球 250,000+ 科学家使用。包含 177 个即用且经过验证的技能,以及 100+ 科学数据库,覆盖生物学、化学、医学和药物发现。兼容 Cursor、Claude Code、Codex、Pi、Antigravity 和开放 Agent Skills 标准。
Turn any AI agent into an AI Scientist. The #1 Agent Skills library for science, used by 250,000+ scientists worldwide. 177 ready-to-use validated skills plus 100+ scientific databases covering biology, chemistry, medicine, and drug discovery. Compatible with Cursor, Claude Code, Codex, Pi, Antigravity, and the open Agent Skills standard.
Python 52,253 4,630 新增 +878
项目简介这是一个科研领域的 Agent Skills 集合,共 177 个技能,覆盖生物信息、化学信息、蛋白质组学、医学影像、机器学习、材料科学等方向。每个技能提供 SKILL.md 说明、工作流、示例或校验,其中数据库技能记录 80 个数据源访问方式。它遵循开放 Agent Skills 标准,可通过 npx、GitHub CLI 或 Agent Plugins 安装,兼容 Claude Code、Codex、Cursor 等主机。
-
6 akitaonrails/ai-memory
为编码 CLI 智能体提供长期记忆,并促进不同 agent 厂商之间交接的解决方案。
Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors
Rust 9,924 718 新增 +178
项目简介ai-memory 为 Claude Code、Codex、Cursor 等编码 agent 提供跨会话、跨工具和跨机器的长期记忆。它以 git 托管的 Markdown wiki 为事实来源,通过生命周期 hooks 自动记录提示、工具调用和会话边界,并在下次继续时生成交接说明。默认零 LLM 调用即可实现捕获、搜索与交接;支持单用户、多用户和团队服务器,并提供 typed claim-once handoffs。
-
7 android/skills
Kotlin 8,428 555 新增 +355
项目简介这是 Android 官方的技能仓库,提供面向 AI 优化的模块化指令和资源,帮助 agent 理解并执行符合 developer.android.com 最佳实践的 Android 开发任务。技能采用开放标准的 SKILL.md 格式,重点覆盖 LLM 表现较弱的工作流。可通过 Android CLI 按单个技能或全部安装,也可作为插件接入 Claude Code、Codex 等;安装后 agent 会在相关任务中自动激活。当前不接受公开贡献。
-
8 stablyai/orca
Orca 是与一组 agent(如 Grok Bot、Claude Code)协作的工作平台。可使用你自己的订阅运行任意编码 agent。提供桌面、移动和远程运行时。
Orca is the place for working with a fleet of agents (e.g. Grok Bot, Claude Code). Run any coding agent with your own subscription. Available on desktop, mobile and remote runtime.
TypeScript 95,419 6,114 新增 +830
项目简介Orca 是一个多 agent 工作台,可在同一界面并行运行 Codex、Claude Code、OpenCode、Pi 等 CLI agent。核心能力包括为每个 agent 创建隔离 git worktree、移动端监控、终端拆分、Design Mode 捕获 UI 元素、GitHub/Linear 原生集成、SSH 远程 worktree,以及对 AI diff 做注释。README 称任何能在终端运行的 CLI agent 都可接入,提供桌面端和移动伴侣,MIT 开源。
-
9 anthropics/claude-code
Claude Code 是一个驻留在终端中的 agentic 编码工具,能够理解你的代码库,通过自然语言命令执行日常任务、解释复杂代码并处理 git 工作流,从而帮助你更快编程。
Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.
TypeScript 151,757 26,877 新增 +179
项目简介Claude Code 是 Anthropic 的终端 agentic 编码工具,可在终端、IDE 或 GitHub 上使用 @claude 交互。它通过自然语言理解代码库,执行常规编码任务、解释代码并处理 git 工作流。仓库提供 macOS、Linux、Windows 和 Homebrew 等安装方式,并包含可扩展自定义命令和 agent 的插件目录;官方文档另说明数据收集与保留政策。
-
10 debpalash/VoiceStudio
VoiceStudio 是开源的、完全本地化的 ElevenLabs 替代品——支持声音克隆、声音设计、视频配音、听写、转录以及 646 种语言的有声书创作。
VoiceStudio is the open-source, fully-local ElevenLabs alternative — voice cloning, voice design, video dubbing, dictation, transcription & audiobook creation in 646 languages.
Python 73,389 8,033 新增 +2,125
项目简介VoiceStudio 提供本地优先的语音处理工作台,功能包括声音克隆与设计、带时间轴的视频配音、听写、转录和有声书批量生成。默认使用 k2-fsa/OmniVoice 引擎,也可选择其他引擎;可在 NVIDIA GPU、Apple Silicon 或 CPU 上运行,并提供 Local API、MCP 和 agent skills 集成。README 强调本地工作流在用户硬件上运行,远程服务为可选;仓库应用部分采用 AGPL-3.0 许可。
-
11 anomalyco/opencode
开源编码智能体。
The open source coding agent.
TypeScript 215,334 28,866 新增 +438
项目简介OpenCode 是一个开源 AI 编码代理,可通过安装脚本或桌面应用使用。仓库内置两个可用 Tab 切换的代理:build 为默认的全权限开发代理,plan 为只读分析与代码探索代理;另有 general 子代理用于复杂搜索和多步骤任务。README 提供安装目录优先级,并提示基于 OpenCode 构建的项目需声明与官方无关。
-
12 ruvnet/ruflo
🌊 最初的 agent harness。部署智能多方协作群体、协调自主工作流,并构建对话式 AI 系统。特性包括自适应记忆、自学习智能、联邦、向量 RAG 集成,以及原生集成 Claude Code / Codex / Hermes 等更多工具。
🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
TypeScript 76,071 9,002 新增 +420
项目简介Ruflo 自述为一个 agent harness,作为 Claude Code 和 Codex 等编码 agent 的执行层,提供 100+ 专用 agent、群体(swarm)协调、自适应记忆、自学习、跨机器联邦、向量记忆/RAG、后台 worker、安全扫描和插件市场等功能。可通过 Claude Code 插件进行轻量安装,或使用 npx ruflo init 完整安装 CLI、MCP 服务、hooks 和 daemon。README 还提供状态、用户指南和验证文档。
-
13 xerj-org/xerj
XERJ 是 AI 搜索数据的新方式。其 autoindex 能力让 agent 在无需像 grep(1974 工具)那样浪费 token 的情况下了解你的数据。一条命令即可索引代码、文档、日志和 PDF,用于搜索、RAG、安全审计和 agent 记忆,token 使用量比 grep 少 40 倍。它经久耐用,可替代 ElasticSearch,节省 10 倍基础设施成本。
XERJ is the new way for AI to search data. Its autoindex capability activates agents to know your data without the token waste of grep (1974 tool). One command indexes code, docs, logs and PDFs for search, RAG, security audits and agent memory, using 40x fewer tokens than grep. It’s durable and can replace ElasticSearch saving you x10 infra costs.
Rust 7,803 416 新增 +1,160
项目简介XERJ 是一个本地 AI 搜索与索引引擎,用一条 autoindex 命令对文件夹中的代码、文档、日志、PDF、SQLite 等自动识别并建立可查询索引。源码经 tree-sitter 解析,保留符号和行号;支持全文、向量、混合搜索,并兼容 Elasticsearch API。它可连接 MCP 暴露搜索、语义检索、记忆存取等工具,适合参考编码、RAG、安全审计与 agent 长期记忆。README 中的对照研究显示其在特定任务中比 grep 驱动的 agent 输出 token 更少。