趋势榜

开源 AI 销售操作系统——自托管 CRM，内置原生 AI Agent 与 WhatsApp（WAHA）。面向以聊天方式开展销售业务的 Kommo、Octadesk、Intercom 的开源替代方案。MCP-ready，多租户，LGPD 合规。Open-source AI sales OS — self-hosted CRM with native AI agents + WhatsApp (WAHA). Open alternative to Kommo, Octadesk & Intercom for any business that sells by chat. MCP-ready, multi-tenant, LGPD.

像构建 API 和微服务一样构建、运行和扩展 AI Agent。Build, run and scale AI agents like API and microservices

一个 R 包，可通过友好、现代的界面创建多种输出格式的样式化表格。An R package to create styled tables in multiple output formats, with a friendly, modern interface.

Wilonity Loader – 一款作弊库，带 spoofer、驱动绕过和未检测注入器，支持 20+ 游戏（RUST、CS2、Valorant、Tarkov、Warzone、PUBG、R6S、ArcRiders、Apex、Roblox、Meccha Chameleon）。具备内核伪装、HWID 清理、AC 绕过（EAC/BE/VG）功能，以及 ESP、自瞄、WH、wallhack、triggerbot、雷达、无后坐力、silent aim、chams、皮肤替换、unlock all、无扩散Wilonity Loader – cheat lib w/ spoofer, driver bypass, undetected injector for 20+ games (RUST, CS2, Valorant, Tarkov, Warzone, PUBG, R6S, ArcRiders, Apex, Roblox, Meccha Chameleon). Kernel spoof, HWID cleaner, AC bypass (EAC/BE/VG). ESP, aimbot, WH, wallhack, triggerbot, radar, no recoil, silent aim, chams, skin changer, unlock all, no spread.

给 linux.do 换装的油猴脚本合集：JetBrains/Darcula IDE 风格、飞书 IM、钉钉 IM 三种外观，仅换皮不碰数据

通过小型 ProgramAsWeights 函数实现英语与 Claudish 之间的相互翻译Translate between English and Claudish with tiny ProgramAsWeights functions.

AI 时代的私人影库

Code World Model 官方仓库。Code World Model Official Repo

使用 ChatGPT 订阅在 Codex 或 Claude Code 中生成 GPT 图像，无需 Images API。Generate GPT images from Codex or Claude Code using a ChatGPT subscription, without the Images API.

TSRX 语言、编译器、工具链以及编辑器集成TSRX language, compilers, tooling, and editor integrations

🦦 ADHDev — Agent 仪表板中枢。在统一仪表板中监控与控制 AI 编程 Agent，自托管开源。🦦 ADHDev — Agent Dashboard Hub. Monitor & control AI coding agents from a single dashboard. Self-hosted, open-source.

在单台 DGX Spark 上运行 Qwen3.8-Flash-Next (180B)，方法是将 51B 的 n-gram embedding 表存放在 NVMe 上Run Qwen3.8-Flash-Next (180B) on a single DGX Spark by keeping its 51B n-gram embedding table on NVMe

面向 AI 编程助手的仿生上下文引擎与神经运行时（Biomimetic Context Engine & Neural Runtime）The Biomimetic Context Engine & Neural Runtime for AI Coding Assistants

2× NVIDIA DGX Spark 上的 GLM-5.3-Flash (NVFP4)——vLLM TP2，262K 上下文，MTP。全球首发部署方案：发现并修复 7 个 day-0 bug，附 sm121 镜像补丁、探针与完整报告。GLM-5.3-Flash (NVFP4) on 2x NVIDIA DGX Spark - vLLM TP2, 262K context, MTP. World-first deploy recipe: 7 day-0 bugs found and fixed, patched sm121 image, probes and full report.

用于高性能终端与原生界面的轻量级 Rust/WASM 原语Lightweight Rust/WASM primitives for high-performance terminal and native surfaces

用于编排 AI 编程 Agent 的委派环境。支持跨 git worktree 管理 Claude、Gemini 与 Codex 会话，集成终端、上下文注入与工作流自动化。A delegation environment for orchestrating AI coding agents. Manage Claude, Gemini, and Codex sessions across git worktrees with integrated terminals, context injection, and workflow automation.

为开发者及其编码 Agent 提供持久化的本地记忆。将命令、输出、错误以及有效修复方案记录到 SQLite，并通过 MCP 提供服务Persistent, local memory for developers and their coding agents. Records commands, output, errors, and the fixes that worked into SQLite and serves them over MCP.

将 LaTeX 或 TeX 数学标记转换为 typst。Translate LaTeX or TeX math markup to typst

面向 VS Code + GitHub Copilot 的开源 AI 研究助手。通过 MCP server 连接 PubMed、OpenAlex、Semantic Scholar、Europe PMC、CrossRef 与 Zotero。定制 Agent 可指导系统综述、学术写作、数据分析与项目管理——全流程符合 ICMJE 标准并保留完整审计轨迹。Open-source AI research assistant for VS Code + GitHub Copilot. Connects to PubMed, OpenAlex, Semantic Scholar, Europe PMC, CrossRef, and Zotero via MCP servers. Custom agents guide systematic reviews, academic writing, data analysis, and project management — all ICMJE-compliant with full audit trails.

自动构建最新版 Overleaf CE Docker 镜像并完整配置为生产就绪环境。Automatically build latest Overleaf CE docker images and fully setup as production ready.

以证据为先的 Agent Skill：对抗性学术新颖性审计——最小先验集合、历史截止点与跨桥感知的先验工作分析。Evidence-first Agent Skill for adversarial scholarly novelty audits — Minimal Prior Sets, historical cutoffs, and bridge-aware prior-work analysis.

单一入口的心理学研究诚信与稿件 Skill，覆盖证据发现、理论、方法、心理测量、统计、学术写作、引文审计、同行评审与投稿发布检查。Single-entry psychology research integrity and manuscript Skill for evidence discovery, theory, methods, psychometrics, statistics, academic writing, citation audit, peer review, and submission release checks.

精选并经核验的资源合集：面向学术与科学写作的 LLM 研究论文、数据集、工具与实现。A curated, verified collection of research papers, datasets, tools, and implementations on LLMs for academic and scientific writing.

华威大学（University of Warwick）本科毕业论文：对 Compulsive Sexual Behaviour Disorder (CSBD) 心理治疗与药物干预的系统综述与 Meta 分析。My final year undergraduate research project at the University of Warwick demonstrating a systematic review and meta-analysis evaluating psychotherapeutic and pharmacological interventions for Compulsive Sexual Behaviour Disorder (CSBD)

期刊查找工具，支持 DGRSDT Class A/B 高级分类与识别，可按出版方、审稿周期、四分位等筛选。Journal Finder with advanced classification and identification compatible with DGRSDT Class A/B, filter by publisher, decision time, quartiles,.etc

有关基于运动信号的自闭症检测 AI 方法综述的补充数据、检索策略与提取文件。Supplementary data, search strategies, and extraction files for the review on AI methods for autism detection from motion signals.

关于 LLM 生成文献综述中知识截止效应研究的精选合集。A curated collection of research on knowledge cutoff effects in LLM-generated literature reviews.

知识截止对快速演化领域中 LLM 生成文献综述的影响Knowledge Cutoff Effects on LLM-Generated Literature Reviews in Rapidly Evolving Fields

UC Berkeley 课程 EDUC 209（Academic Writing Support Group）的教学大纲模板与资源。该课程我开设多年，部分同行觉得有用，现以 Markdown 文件公开，便于导入 Google Docs、转换为 HTML 或跨平台编辑使用。Syllabus template and resources for my longstanding UC Berkeley course EDUC 209: Academic Writing Support Group. Some colleagues have found these useful, so here they are rendered from simple Markdown files: a human-readable filetype that you can import into Google Docs, convert into HTML files, or otherwise edit and use across various platforms.

LLM 的知识截止会使文献综述过时或不完整；RAG 与时序感知检索可提升时效性，但无法完全解决引用、来源选择与时间推理问题。LLM knowledge cutoffs can make literature reviews outdated or incomplete, while RAG and time-aware retrieval can improve freshness but do not fully solve citation, source-selection, and temporal reasoning problems.

证据合成自动化框架——PRISMA 流水线：检索、去重、AI 筛选、Meta 分析（D-L 随机效应）、PRISMA 流程图与森林图。Evidence Synthesis Automation Framework — PRISMA pipeline: search, dedup, AI screening, meta-analysis (D-L random effects), PRISMA/forest plots.

Episteme：可复现、无外部依赖的系统综述与 Meta 分析引擎（PubMed → PRISMA → Meta 分析 → SVG 图表，一键完成）。无需付费 API，无需 numpy，无需 R。Episteme: reproducible, dependency-free systematic-review & meta-analysis engine (PubMed -> PRISMA -> meta-analysis -> SVG figures in one command). No paid APIs, no numpy, no R.

由 5 个 AI Agent 协同的 pipeline，自动回答研究问题。可类比为聘请 5 位专家：规划员、资料员、记录员、事实核查员与撰稿人。A pipeline of 5 AI agents that work together to answer a research question autonomously. Think of it like hiring 5 specialists: a planner, a librarian, a note-taker, a fact-checker, and a writer.

Journal Finder V7.0（期刊匹配工具 V7.0）。Journal Finder V7.0

Journal Finder V6.0（期刊匹配工具 V6.0）。Journal Finder V6.0

Journal Finder V4.0（期刊匹配工具 V4.0）。journal Finder V4.0

一套自动化 Python 流水线，用于儿童营养不良数据的可复现统计分析，涵盖数据清洗、EDA、多重共线性评估、逻辑回归、模型诊断、出版级表格、图表与自动化报告。An automated Python pipeline for reproducible statistical analysis of child malnutrition data, including data cleaning, EDA, multicollinearity assessment, logistic regression, model diagnostics, publication tables, figures, and automated reporting.

使用 SPSS 统计分析，分析 AI 语言模型对沟通、创造力与信息流影响的研究项目。Research project analyzing the impact of AI language models on communication, creativity, and information flow using SPSS statistical analysis.

文章《Approach-avoidance tendencies and physical activity level: A systematic review and meta-analysis》（趋近-回避倾向与体力活动水平：系统综述与 Meta 分析）的代码、数据及相关材料。Codedata, and material for the article: "Approach-avoidance tendencies and physical activity level: A systematic review and meta-analysis"

在 FlyRank AI 进行的机器学习工程实习工作，包含 ML 实验、研究问题、AI workflow 与应用机器学习项目。Machine Learning Engineering Internship work at FlyRank AI, including ML experiments, research questions, AI workflows, and applied machine learning projects.