Havuz

A multimodal RAG dashboard and interactive 3D Knowledge Graph. Process documents locally with Ollama or via Cloud APIs, powered by LightRAG,

NTK — 基于信息密度路由的多智能体框架。用最少的 token 做最多的事。

Official code for the paper: "Multi-User Large Language Model Agents"

Research-backed Harness Engineering study repo with a runnable Python MVP built around Plan -> Work -> Review.

WikiMind is a personal LLM-powered knowledge OS. Feed it articles, PDFs, YouTube videos, podcasts, or papers — it compiles them into a struc

qMLX: Custom inference engine for Qwen 3.5 122B on Apple Silicon, extending MLX with hybrid attention support, SSD-backed KV cache, and RYS

国内公募基金数据采集与AI辅助分析工具 — 全栈FastAPI+React，集成AKShare多源数据、风险评分引擎、回测验证、LLM可选分析

NetOpsBench: Open Arena for Agentic NetOps in AI Infrastructure

Odylith stops coding agents from confidently doing the wrong thing

Terminal-native agent for clinical workflows — prior auth, coding audit, and documentation, with reviewable output

Open-source vector database and sub-microsecond key-value store in one Go engine — embed it as a library, run it standalone, or replicate it

Behavioral-discipline skill turning an AI agent into a methodical, non-robotic bug-bounty & pentest hunter - 42 always-on rules: scope, cove

Simple Long Horizon Agent - A simple yet effective AI agent for learning, experimentation, and long horizon work.

Multiple agents, one source of truth: a Planner/Coder/Reviewer demo of how a shared Statewave subject prevents parallel-agent context collis

TERSE - a hierarchical state language for humans and AI agents.

Audit MCP servers, skills, and CLAUDE.md bloat eating your Claude Code context window.

Interactive AI tutor built as a Claude Code skill — teaches with spaced repetition, personalized project examples, and a visual progress das

Lightweight RAG provenance middleware. Verifies every claim in an LLM response is grounded in a retrieved source - without an LLM call.

Claude Code skill: Multi-agent adversarial review panel — 4-6 AI reviewers debate your code/plans, then a supreme judge delivers the verdict

MCP server that lets AI assistants fetch content from authenticated web pages.

Convert long AI conversations into portable conversation state graphs for LLM handoffs.

把高质量播客和长文 RSS 自动转成 karpathy-claude-wiki 兼容的 source-summary 页面 / Turn high-signal podcasts and long-form RSS into karpathy-claude-wiki compat

Concurrent Go CLI that detects hidden prompt-injection content in documents — invisible Unicode, homoglyphs, hidden HTML, and known trigger

Autonomous AI dev team — 9 Claude agents run as an agile team: plan sprints, write code, review, test, deploy. Runs 24/7 via cron.

RoR engine that enables natural language queries over your database using RAG and LLMs.

🚀 多Agent协作智能知识库系统 — 五层可编排Agent架构 + RAG全链路 + 统一ToolRegistry，Spring Boot + FastAPI + React，核心亮点是 五层可编排 Agent 架构——Interface、Orchestrator、Tool、M

Uni-ViGU: Towards Unified Video Generation and Understanding via A Diffusion-Based Video Generator

Agentic AI development loop that splits Design, Code, Review, and Test into isolated roles with deterministic guardrails

Free, open-source AI coding agent & multi-agent teams for macOS — runs 100% on-device via Ollama or LM Studio. Autovisor, an autonomous Supe

Deterministic agentic-style multi-hop RAG at research-SOTA parity on commodity LLM APIs — no GPU, proof tree per answer.

Validate and retry LLM outputs for ruby_llm. Describe the JSON response you expect, fall back to a stronger model when the cheaper one fails

Use Claude Code's UI with OpenAI GPT-5.6 (Sol/Terra/Luna) as the engine — powered by your ChatGPT Plus/Pro subscription via a local Codex pr

The networked cross-session for coding agents — @mention another team's AI so yours and theirs integrate directly across sessions, machines,

A scheduled, agent-driven software factory. Keep a plain JSON backlog; Forgeo runs your coding agent (Claude Code, Codex, opencode) on each

Claude Code plugin for building LLM-maintained Obsidian wikis from raw research — compile, query, lint, and evolve your personal knowledge b

إطار مفتوح المصدر لتكييف مساعدات الذكاء الاصطناعي وسير العمل القانوني مع البيئة القانونية السعودية. Open-source framework for adapting AI as

CLAUDE.md + AGENTS.md 12-rule behavior file for Claude Code, Codex, Cursor, Hermes — drop-in, ~700 tokens, cuts AI coding mistakes ~40%→~3%

The hosted share target for crit - Your feedback loop with the agent.

An open-source AI workspace for knowledge workers. Bring your own AI subscriptions and use the providers, harnesses, and models you choose t

Generates executable Proof-of-Concept for any bug in any project. AI agents discover and reproduce vulnerabilities — verified, not hallucina

Multiplayer infrastructure for AI agents

Complete memory and context persistence system for OpenClaw AI agents

Typed structured extraction from PDFs, photos & text via any LLM — Instructor for Swift, document-first

Macaw — on-device AI agent for macOS. 2.7B LFM2.5 fine-tune, MLX 4-bit, 97 macOS tools, no cloud.

Lightweight, dependency-free logging and analytics for LLM API calls — track tokens, cost, and latency to a plain JSONL file.

A runnable reference implementation of multi-agent constraint propagation using Caura. 5 specialists share memory via MCP so decisions made

一个常驻桌面的 AI 小伙伴。悬浮球形态，轻量不打扰，支持对话、文件处理与轻记录，具备 Agent Loop 与按天记忆能力。Electron + TypeScript 构建，支持 macOS & Windows。

Any Model for VS Code Copilot

Run Claude Code on any LLM provider without losing your setup, local-first proxy with verified profiles, OAuth device-flow login, failover,

Transparent, transport-layer semantic cache for LLM API calls, powered by Redis 8 Vector Sets.