A context compression toolkit that cuts LLM token usage by 60–95% — pluggable as a library, proxy, or MCP server for tool outputs, logs, files, and RAG chunks. 专为 LLM 设计的上下文压缩工具包,可作为库、代理或 MCP Server 接入,将工具输出、日志与 RAG 片段的 token 消耗削减 60–95%,答案质量几乎无损。
NousResearch's adaptive agent framework — built to grow with your workflows and evolve through use, rather than requiring exhaustive upfront configuration. NousResearch 推出的自适应 Agent 框架,随使用深入能力持续迭代,无需预先繁琐配置,自然融入实际工作流之中。
An agent harness that bakes skills, instincts, memory, and security into Claude Code, Codex, Cursor, and other AI coding assistants — optimizing their performance from the ground up. 将技能、直觉、记忆与安全机制统一注入 Claude Code、Codex、Cursor 等 AI 编码工具的智能体优化框架,从底层提升模型实战表现。
The definitive self-study roadmap for aspiring software engineers — a battle-tested curriculum covering every CS fundamental needed to ace the technical interview. 面向目标工程师的完整自学路线图,久经实战检验的 CS 核心课程涵盖算法、数据结构等面试必备知识体系。
An end-to-end local platform for Live2D AI companions — supports any LLM with hands-free voice interaction, voice interruption, and cross-platform streaming. 全本地的 Live2D AI 虚拟主播框架,接入任意 LLM 并支持语音免手动交互、打断说话与多平台直播推流。
A Windows companion suite for OpenClaw — includes a System Tray app, shared library, and PowerToys Command Palette extension for a polished native desktop experience. OpenClaw 的 Windows 伴侣套件,集成系统托盘应用、共享库与 PowerToys 命令面板扩展,打造流畅的原生桌面体验。
GitHub's official toolkit for Spec-Driven Development — helps teams define and validate software specs before writing a single line of implementation code. GitHub 官方出品的规格驱动开发工具包,帮助团队在动笔编码前先对齐并验证软件规格,从源头提升工程质量。
A graph-based OSINT investigation platform — visually map, link, and query entities for security analysts and digital investigators. 面向安全分析师与数字调查人员的图式 OSINT 平台,以可视化节点图谱呈现、关联并查询实体关系。
Aqua Security's comprehensive vulnerability scanner — finds CVEs, misconfigs, secrets, and SBOM across containers, Kubernetes clusters, code repositories, and cloud configs. Aqua Security 出品的全栈漏洞扫描器,覆盖容器、K8s、代码仓库与云配置,一键检出 CVE、错误配置、密钥与 SBOM。
An open-source NotebookLM alternative with more control and flexibility — supports custom LLM backends and deeper notebook management workflows. NotebookLM 的开源替代品,后端可自由配置、工作流更灵活,完整掌控你的 AI 知识笔记系统。