仓库/Skill 库

有界研究 Agent：plan → retrieve → critique，具备步骤预算、执行轨迹、引用追溯与自由路径 UI，无需 API key。Bounded research agent: plan → retrieve → critique, step budgets, traces, citations, free-path UI. No API key.

在 Claude 中强制以引用优先的研究流程，通过结构化五阶段工作流验证事实、识别不确定性并消除幻觉。Enforce citation-first research with Claude to verify facts, identify uncertainty, and eliminate hallucinations through a structured five-phase workflow.

🔧 通过灵活的 prompt 模板与任务规范项目简化 AI Agent，实现高效的指令与记忆管理。🔧 Streamline AI agents with this flexible prompt template and task specification project for effective instruction and memory management.

Subconscious Memory 2026：基于混合语义搜索与 MCP 的 AI 编码上下文引擎。Subconscious Memory 2026: AI Coding Context Engine with Hybrid Semantic Search & MCP

📰 Spec-driven daily briefings on Claude Code cloud Routines — git as the only state, zero servers, zero API keys | 用 Claude Code Routines 每天自动生成主题简报，git 即全部状态，零服务器零密钥

可复现的研究流水线：将基于 Agent 的模型产出转化为有证据支撑的 LLM 高管报告。A reproducible research pipeline for converting agent-based model artifacts into evidence-backed LLM executive reports.

面向编码 Agent 的本地运行时，以合并质量为优化目标：由仓库自适应常驻模型生成补丁，大模型进行验证。单进程同时支持 Anthropic + OpenAI API，天生离线运行。A local coding-agent runtime that optimises for merge quality: a repo-adapted resident model generates patches, a large model verifies them. Anthropic + OpenAI APIs from one process, offline by construction.

SwiftUI 原生 LaTeX 数学渲染，对 LLM 输出具有鲁棒性——支持分隔符归一化、流式渲染与自动主题适配SwiftUI-native LaTeX math rendering robust to LLM output — delimiter normalization, streaming support, and automatic theming

将 OpenCode 转变为可复现的研究工具Turn OpenCode into a reproducible research instrument.

大语言模型（LLM）是一种人工智能（AI）程序，能够识别并生成文本，以及执行其他任务。A large language model (LLM) is a type of artificial intelligence (AI) program that can recognize and generate text, among other tasks.

在双 RTX 3090 GPU 上以 262K 上下文长度并发流运行 Qwen3.6-27B，开发工作已迁移至 club-3090。Run Qwen3.6-27B at 262K context with concurrent streams on dual RTX 3090 GPUs. Development moved to club-3090.

用于生成文献综述的全栈系统，基于导入的数据构建，用户可自由选择 LLM。A full stack system use to generate literature review base on dataa imported, users can choose between llms.

基于 TypeScript 的独立框架，通过统一原语 Component 组合安全的个人 AI agent：manifest-first 发现、能力授予、正交模型路由与信息流控制。Standalone TypeScript framework for composing secure personal AI agents from one universal primitive — the Component. Manifest-first discovery, capability grants, orthogonal model routing, and information-flow control.

针对 PX4 飞控软件的覆盖率引导事务语义审计研究 artifactResearch artifact for coverage-guided transactional semantic auditing of PX4 flight-control software.

基于多 Agent LLM 工作流的智能 AI 学术研究助手，可自动化完成学术论文发现、验证、文献分析、对比矩阵（含表格）、RAG 驱动的研究对话，并产出可直接用于稿件的文献综述An agentic AI academic research assistant that automates academic paper discovery, validation, literature analysis, comparison matrices with table, RAG-powered research chat, and manuscript-ready literature reviews using multi-agent LLM workflows.

完全在浏览器中通过 WebAssembly 实现文档转换、检查、OCR 与翻译。保留版式的翻译，全程本地、离线优先。Convert, inspect, OCR, and translate any document entirely in the browser via WebAssembly. Layout-preserving translation, fully local, offline-first.

面向 AI 饮食评估研究的 3 周入门计划：文献综述、分割、LLM 与 RAG。3-week onboarding plan for AI dietary assessment research: literature review, segmentation, LLMs, and RAG

使用支持 macOS 和 Windows 平台的开源工具高效构建与定制机器学习模型Build and customize machine learning models efficiently with an open-source tool that supports macOS and Windows platforms.

为 LLM Agent 提供探索与调用任意 OpenAPI/Swagger/Postman API 能力的 MCP server，支持搜索、检查与调用。MCP server that gives LLM agents superpowers to explore and call any OpenAPI/Swagger/Postman API — search, inspect, invoke.

探索高达 400B 参数的先进稀疏 Mixture-of-Experts 模型，采用创新训练技术实现卓越性能🚀 Explore advanced sparse Mixture-of-Experts models with up to 400B parameters, featuring innovative training techniques for superior performance.

为 Claude 提供跨会话持久记忆——将 TriliumNext 转化为持久化、图结构化第二大脑的 MCP server。MIT 许可。Give Claude a memory that survives the conversation - an MCP server that turns TriliumNext into a persistent, graph-structured second brain. MIT.

MAI-Code 模型的官方仓库，用于发布 MAI-Code 模型版本与更新，并通过 issue 与反馈与开发者社区互动。Official repo for MAI-Code models, where we publish MAI-Code model releases and updates and engage with developer community on issues and feedback.

Microwave Method：agent 工厂 + 受治理的 LLM wiki。每个 agent 一次性重型处理，每个 feature 轻量处理。上下文只"烹饪"一次，按缓存价"回热"。Microwave Method: agent factory + governed LLM wiki. Heavy pass once per agent, light pass per feature. Context is cooked once, reheated at cache price.

使用纯 C 从零构建 LLM 推理引擎，不依赖任何框架。Build an LLM inference engine from scratch in pure C with no frameworks.

面向 Microsoft Word 的 AI 驱动学术校对工具，支持修订痕迹、审稿意见与写作润色，兼容 macOS 与 Windows。AI-powered academic proofreading for Microsoft Word — tracked changes, reviewer comments, and writing polish on macOS & Windows.

Portable Agent Plugins 1.0 skill，支持 STAR 简历撰写、指标推断、XeLaTeX/PDF 排版、视觉 QA 与面试准备。Portable Agent Plugins 1.0 skill for STAR resume writing, metric inference, XeLaTeX/PDF typesetting, visual QA, and interview preparation.

对 LLM 文献筛选方法的受控双研究对比——RAG 与 zero-shot、自适应与被动选文——并以真实系统综述的真值标注进行评估。Controlled two-study comparison of LLM-based literature screening methods — RAG vs. zero-shot, and adaptive vs. passive paper selection — evaluated against real systematic review ground truth.

借助这款自托管、本地优先的应用，跨多个账户与多种货币追踪收入和支出。Track income and expenses across multiple accounts and currencies with this self-hosted, local-first application.

探索由 19,846 篇量子计算论文组成的 AI 原生知识库，面向搜索、趋势分析与 LLM 友好的研究工作流Explore an AI-native knowledge base of 19,846 quantum computing papers, built for search, trend analysis, and LLM-ready research workflows

BYOK AI 科研工作台——提供科学写作、可编辑图表、PPT 生成与参考文献管理，支持 12+ 模型提供商。BYOK AI research workbench — scientific writing, editable figures, PPT generation, and reference management. Supports 12+ model providers.

🐺 创新型 LLM 驱动 12 人狼人杀模拟，针对直播优化并支持 OBS 无缝集成🐺 Experience an innovative, LLM-driven Werewolf simulation for 12 players, optimized for live streaming and designed for seamless integration with OBS.

通过单一严格的 YAML 清单统一管理本地 LLM 运行时——支持状态查看、健康检查、启动、停止，基于原生 systemd 与 Docker。Manage local LLM runtimes from one strict YAML manifest — status, doctor, start, stop over native systemd and Docker.

本地使用 AI 总结视频，从讲座、会议和教程中提取视觉亮点与文本摘要，基于 RTX 40 系列 GPU 的 CUDA 加速Summarize videos locally with AI, extracting visual highlights and text summaries from lectures, meetings, and tutorials using CUDA on RTX 40-series GPUs

每日提示词，激发 nano banana 生成灵感。Daily prompts to inspire nano banana generation.

为 AI agent 提供持久且安全约束的浏览器控制——使用语义化 CLI 替代 Playwright 脚本或屏幕坐标。通过统一接口，背后使用 macOS WKWebView 与 Linux ChromiumPersistent, safety-enforced browser control for AI agents — semantic CLI instead of Playwright scripts or screen coordinates. macOS WKWebView + Linux Chromium behind one contract.

AI 时代的工具联盟：可追溯的知识（Nexus·Archon）、可问责的审查与治理（Arbiter·Observer）、可理解的变更质量（Adept·Probe）。AI 构建，你来理解。An alliance of tools for the AI era: grounded knowledge (Nexus·Archon), accountable review & governance (Arbiter·Observer), and comprehension & mutation quality (Adept·Probe). AI builds it — you understand it.