Custom agents

LunaScope 0.5.0 — completely rebuilt Windows-first local AI agent workbench with UltraNote.

A股热门板块多智能体量化分析系统 | 基于FinGenius架构 | ETF50+60只个股扫描 | AkShare数据源免费

Bookkeeping for coding agents. Hand it an issue; it claims the work, writes the run's mission list (item / status / dispatched / result), and runs it. Resumable on disk. Six runtimes — Claude Code, Codex, Cursor, opencode, Kimi Code, and Grok CLI — across GitHub, GitLab, and Gitea.

Prudence, a custom harness powered by Microsoft Agent Framework. Includes a "counsel and craft" mode which delegates executive function to Microsoft Foundry and task completion to local models.

Evoflux is an open-source, local-first workspace where AI agents build software, conduct deep research, automate browser tasks, and collaborate in parallel. Connect any model, keep control of your workspace and data, and take complex work from idea to completion—all in one place.

Always-on Windows voice assistant with a 36-tool agentic layer (Claude). Local wake word + STT, streamed TTS, sandboxed code execution, vision/security, phone + Discord clients.

Multi-agent system that unifies heterogeneous dental data (CBCT, STL, clinical reports, photos) into a patient Digital Twin built on Gaussian Splatting, with per-region clinical attributes and longitudinal series.

Amazon Bedrock AgentCore enterprise platform accelerator with AWS CDK, Terraform organization guardrails, MCP/A2A agents, security, memory, and observability.

An NLP-powered, Agentic AI approach to Perfumerie. Create custom fragrances using poetry.

Caduceus is a Unix daemon, shipped as a Hermes plugin, that polls GitHub for labeled issues, runs your AI harness against them in isolated worktrees.

An in-QGIS agentic chat assistant. Type a request in plain language and an LLM agent carries it out inside the running QGIS session If QGIS can do it, the agent can do it: anything you could script in the Python console or run from a toolbox, it reaches the same way.

An open, vendor-neutral organizational model for building and transforming AI-first companies with organizational intelligence, governed autonomy, and controlled execution.

A modular AI reasoning framework for structured thinking, expert orchestration, verification, and high-quality problem solving.

Simple Agent orchestration

Typed Determinism for Probabilistic AI

A modular, extensible toolkit for red teaming large language models and NLP systems — covering adversarial text attacks (character, word, sentence, semantic), jailbreak evaluation via JailbreakBench, and prompt injection — with pluggable model targets, automated judges, and clean reporting. Built for researchers and AI safety practitioners.

AI-first engineering toolkit that turns coding agents (Claude Code, Codex, Cursor, Windsurf, Gemini CLI, Aider) into Senior mobile engineers — agents, skills, workflows, checklists & standards for iOS/Android/Cross-Platform.

Private, local-first AI assistant with cited RAG, Ollama, a sandboxed agent, and scoped LAN pairing.

Cross-LLM skills for personal planning, research writing, mathematical review, and assistant development.

MLHub is a collection of impactful machine learning projects designed for learners and enthusiasts in the field of data science. Our goal is to provide accessible and hands-on experiences that help individuals understand the fundamentals of machine learning and data analysis.

Retrieval-augmented conversational AI on Cloudflare's edge: SSE streaming, Vectorize RAG, KV-backed memory, and an in-worker abuse guard. The deployed face of my multi-agent, DAG-orchestrated system.

This repo holds the memory engine of EVE, an AI chatbot for erotic conversation

AI-powered company chatbot for answering FAQs, support queries, and business information.

This repo holds the reasoning algorithms engine of EVE, an AI chatbot for erotic conversation

The open cybersecurity harness - by the Swiss Applied AI Cybersecurity Research Lab.

dsh plugin: run work as a small crew of role agents (product manager, researcher, architect, engineer, QA, code reviewer, security reviewer, doc reviewer) that share work through files on disk

CatchBench: a benchmark for auditing agent failures across the full lifecycle.

A team of little AI helpers that live on your computer. They remember you, they do real work, and your data never leaves your machine.

Security-first AI platform on Kubernetes — FastAPI + Next.js chatbot with RAG, self-hosted model serving (vLLM), observability, guardrails, and model gateway, deployed on Azure AKS.

Enterprise AI Business Intelligence platform featuring semantic search, knowledge engine, persistent vector storage, and production-oriented architecture built with FastAPI, FAISS, and Python.

Self-hosted RAG system with Ollama embeddings and Supabase/pgvector. 100% local, 100% free.

This repository contains SupplyChain-AI: An AI-powered generative chatbot leveraging LLMs, RAG architecture, and predictive analytics to answer conversation queries in supply chain management.

Privacy-first, local-first AI meeting assistant for macOS & Windows — live transcription, an in-meeting copilot grounded in your own documents, and summaries. Everything runs on your machine. No cloud, no telemetry.

Turn scattered documents into a knowledge base that maintains a position over time. One append-only ledger of evidence-bearing assertions; the RAG index, knowledge graph, OKF wiki and change report are projections of it.

photon_db is high performance vector database built using Rust with Python bindings

Yu-Gi-Oh! OCG AI ruling and judge assistant｜游戏王AI裁定与规则问答｜遊戯王AI裁定・ルール質問 — official Q&A retrieval, RAG, multilingual.

Open-source AI travel planning workspace for evidence-backed itineraries, maps, reports, and printable PDFs.

DSH插件-模型余额显示功能

Automate video, note, product-link and scheduled publishing across social media and e-commerce platforms.

DeepSeek Harness (DSH) plugin for self-improving AI agents: continual learning, persistent memory, cross-session knowledge, review-and-refine workflows, and automatic rollback.

Google Calendar, Gmail, Tasks, Drive & Contacts MCP server for Pi. 26 tools, single binary, OAuth login.

Official MCP server for AllMCPs.com — submit MCP servers to the directory directly from your AI agent.

Terminal-first AI coding harness for CachyOS/Arch + Kitty + Helix + omp (Rust, low-RAM).

郭帆导演 Skills — 你的科幻短剧脚本创作全流程 AIskills

A bounded math reasoning agent with Lean 4 verification — conjecture, proof, verified conclusion

🔮 Lucy — a screen-aware AI voice agent for Android. She sees your screen, understands it with Gemini, and taps and types for you.

闲鱼 / Goofish AI 自动回复本地控制台：商品专属策略、自动发货参考、Cookie 同步辅助、Ollama 本地模型和 Windows 托管脚本。

15-minute beginner coach for Guan Dan, a Chinese four-player partnership card game — deterministic rules, fair AI, memory drills and replays.

让任意 Harness（ZCode/Claude Code/Codex…）调用 OpenCode 作为子 Agent：并行派活、续跑迭代、模型自由——不绑定主 Agent 供应商，可混用 DeepSeek/GLM/Kimi/免费档/Claude 等任意套餐，节省高级模型 token（AI skill + CLI）

🍟 hermes-fry-cards — 飞书流式卡片插件

An execution harness for gptel-agent that adds reliable, autonomous coding-agent behavior.

Self-evolving memory + skill lifecycle for DeepSeek Harness — durable cross-session memory with zero-token deterministic recall, tiered approval, reinforcement learning from repetition, and anti-bloat convergence for both skills and memory.

把视频号、B站、小红书或抖音链接发给 Hermes，在 Mac 本地完成下载、整理和逐字稿；支持视频号、B站、抖音博主批量抓取。

Project page for Embodied Agent Interface: LLM agent benchmark for embodied decision making. NeurIPS 2024 Oral.

OpenCode plugin: cross-session memory for recurring tool-call failures. Detects repeating bash/read/edit/grep failures, promotes them into enforced gates — remind first, hard-block on same-session repeat. TypeScript + Bun, no build step.

Open-source AI terminal for China A-shares — stitches the best open-source quant projects (TradingAgents, Kronos, Chronos-Bolt, XGBoost) into one pipeline: multi-agent analysis → predict → verify → WeChat bot. 开源A股AI投研终端:缝合市面最强开源量化项目,预测-验证闭环+微信机器人

Local-first AI company operating system with a Codex lead, Claude and OpenCode workers, and a live pixel office.

AI 驱动的多 Agent 自动化渗透测试平台：LLM 协调员+监察员编排专项子 Agent，授权边界内自动完成侦察→漏洞检测→利用验证→取证报告。Web 渗透 + C2 双作业线，45+ 原生工具，Docker 一键部署

A local-first multi-agent status island for Windows

Unified fleet-wide agent memory system for DeepSeek Harness — shared Obsidian Vault for dsh, Codex, Claude Code & Hermes with dependency-free Python FTS5 core.

A multi-agent red-team tribunal for your hardest engineering decisions: AI CLIs from different vendors debate in structured rounds and return adjudicated verdicts that preserve dissent. A methodology, not a framework.

Multi-LLM development harness — local-first, test-gated, honest by construction. Requirements→spec→plan→build→review→release with real verification gates; operable by humans, CLI, desktop, or other agents via MCP.

Teach once, replay anywhere — harness-agnostic Teach-by-Demonstration engine (Record → Compile → Replay → Verify) with OpenMausBot and DeepSeek Harness integrations.

Import high-quality conversations into Claude Code and provide CLAUDE.md recognition, add basic features like the "/init" command, and help speed up migration progress | 高质量导入Claude Code中的对话，并提供CLAUDE.md识别，添加"/init"命令等基础功能，帮助快速迁移进度

Review your AI agent’s security configuration before deployment. Emphasizes local, deterministic setup under Apache-2.0 license.

Personal engineering skill framework for agentic AI development(Inspired by Karpathy) — fork and adapt

Python bridge between MAX Bot API and Hermes Agent. Webhook, Docker, systemd.

AI-powered interview preparation platform with mock interviews, performance analysis, and personalized feedback.

Native MAX Bot API platform plugin for Hermes Agent. Direct integration without bridge.

The Small Language Model Chatbot is a Python-based interactive chatbot built using Hugging Face's distilGPT-2 model. The chatbot is fun, lightweight, and ideal for demonstrating the capabilities of small-scale language models. It includes interactive widgets for user input and a visual element that tracks and displays .

GoRAG is a production-ready, high-performance RAG (Retrieval-Augmented Generation) framework built entirely in Go. Designed for enterprise scalability, it seamlessly connects your internal data to the most powerful LLMs with zero Python dependencies.

MoziAI-35B: Local open-source financial AI LLM. 35B MoE model compressed to 15.5GB via MoziSmartBit quantization. 256K context, vision, tool calling, 140+ tok/s on consumer GPU. Based on Qwen3.6-35B-A3B.

Local, offline backup & viewer for your AI chat history (Gemini, ChatGPT, Claude)

Show an AI coding assistant only the code that matters for the change it's making. Measures whether it actually works on your repo.

⚡ AI-powered IDE with RAG pipeline (pgvector) and multi-agent DAG orchestration via NATS

A complete, verified guide to becoming an AI Product Manager and passing AI PM interviews: LLM and ML foundations, evals, RAG, agents, product craft, 7 industry domains, sourced case studies, a 160+ question interview bank, and career resources. Every fact sourced, no hallucinations.

AISMM (AI-driven Software Meta-Model) is a structured, machine-readable and human-readable way to describe a software product as a complete system.

-

Adaptive Graph-RAG extending Microsoft's GraphRAG with query-complexity routing, incremental graph updates, and self-verified retrieval to reduce cost and hallucination.

Air-gapped enterprise RAG workstation — local inference, zero cloud egress. Built with Microsoft Foundry Local.

Lightweight vector search engine on top of Valkey

A model-facing knowledge canon for incentives, interdependence, equilibrium, institutional design, strategic failure, and long-run governance.

Shongjog — Offline Bangla Multi-Hazard Emergency Companion Powered by Gemma 4. Emergency AI for 12 disaster types. Gemma 4 on-device works offline, gets better online, and meshes phone-to-phone.

Open-source AI intelligence platform for private equity — cited RAG chat over your data room, portfolio analytics, AI model builder, meeting transcription & LP inbox. Self-hosted in your own AWS (full Terraform). AGPL-3.0 alternative to Rogo / Hebbia.

Sovereign Agentic OS — self-hostable, EU-sovereign agentic data platform on Kubernetes (pre-beta). One Helm chart, runs in your own cluster.

Karpathy-style LLM Wikis for the Hermes Agent — persistent, compounding knowledge bases curated by AI agents

面向外贸团队的证据约束客户研究、产品推荐、市场对比、开发材料与可编辑演示文稿 Agent，保留人工审核与独立审批。

Stop searching your documents. Start asking them. Transform Paperless-ngx into a conversational AI assistant using Agentic RAG and the Model Context Protocol (MCP).

Evaluating RAG robustness against corpus poisoning and query-side noise using retrieval and answer-quality metrics.

Palace memory kernel (Go) — local-primary hierarchical FS palace, hybrid search, temporal lite; not product Memory GA

High-performance MCP server for browser automation, web capture, and content extraction. Rust-powered CDP client for AI agents.

LLM Framework for LLMs

A free, open-source LLM security gateway. Drop it between your app and any LLM provider to get rate limiting, audit logging, token tracking, and Prometheus metrics with no code changes to your app.

A modular, extensible Rust-based server providing short-term, long-term, and latent memory services, a chat endpoint with a BaseAgent + Sentience DSL, and integration with ChromaDB and LLM services.

Lightweight (and therefore powerful) Rust agent framework inspired by Browser Use's Agent SDK

UI interface for Kitaru - durable execution for AI agents, built on ZenML

An agent native CLI wrapper around delve to verify or debug golang applications No MCP

Shadow agents for autonomous code work - a local-first agentic work orchestrator

AgentPin agent pinning protocol, part of the Symbiont Agent Trust Stack

Multi-user AI Agent Core via asyncpg JSONB GIN Index, In-Memory KV Cache & Hybrid Static-Dynamic Role System.