Custom agents

HY-AI4Science

AYGENT is an Open Source AI Agent Harness with built-in security, memory, cron, and parallel threads.

Next-gen utility API for AI agents & developers: Prompt Injection Guardrails, LLM context sanitization, web scraping, and data validation.

A dev guardrail for fast moving teams

The ReAct loop, tools, context engineering, memory, planning, MCP, evals and security — built from scratch in dependency-free TypeScript.

Official extension collection for posoco

ECHO-CLAW is an AI Remote Viewing training framework built on a Monitor–Viewer architecture, blind-target separation, structured memory, and cumulative learning. It combines theoretical design with practical OpenClaw-based workflow for disciplined, target-blind AI session work and long-term experimental development.

35 pinned source snapshots for building AI agents: coding runtimes, SDKs, MCP, UI and skills. Bilingual reading routes and original upstream links.

ReAct agent loop (Thought, Action, Observation) in synchronous Rust with no async runtime or I/O, so it builds for wasm32 and runs in browsers, edge workers and plugins.

The open standard for AI agent organizations. Define, validate, and deploy AI teams with ranked roles, communication protocols, and escalation rules.

🏭 AI 工程化交付系统 — 把完整工程团队的工序编进一条多 Agent 管线：14 角色 · 三道验证闸门 · EXIT_CODE 指纹 · 定向回环 · 三层记忆。实战产出：AgentGate（127 项测试）。

基于 langgraph 图工作流，dsh cordis 插件思想框架的通用agent内核

ACSE — a sovereign cognitive agent built on SQCIG (Schematized Quadrupole Causal Ignition Graph): an A-D-O-G cognitive quadrupole wired as a living causal graph, in a single Zig binary. Every claim measured, never declared.

Run a fleet of personal AI agents — each a complete, self-contained git repo — behind one connector and a web client.

Run pi, the open-source coding agent, inside Restate durable execution: crash-proof agent turns, steering, human approvals and multi-agent calls

Ep. 19 — Code Interpreter with Microsoft Foundry: charts and PDF reports

Open Benchmarks and reproducible evaluation harnesses for Shadow AI Agent Discovery & Inventory

Project-scoped rule memory for coding agents.

A minimal context kernel for AI agents

Local-first governance layer for AI agents. Policy engine, audit trail, rollback, cost control. All offline.

自托管的开源多模型 AI 智能体应用:ReAct 循环 + 计划执行 + 长期记忆 + 上下文压缩 + 子代理 + 人工审批 + MCP 扩展 + 流式思考 + Trace 回放 + 内置评估框架(25 任务 100% 通过)。14 个内置工具,五层安全防护,Flask + SQLite 轻量架构,双击即可部署,数据完全私有。

Composable multi-agent framework for GitHub Copilot — persona grafting, cognitive identity caching, and agent fusion for VS Code.

Run a company of AI agents out of one git repository — any craft, not just code. The team, the process, the history, the cost — all files: clone it and everything travels. Gates that say what enforces them, doors that refuse instead of hoping, evidence on every claim.

Reliable, embeddable TypeScript agent runtime with durable runs, tool safety, recovery, evidence, and bounded multi-agent coordination.

Local-first collaboration hub for AI agents - shared knowledge base, user memory, async messaging, skill market and usage analytics behind one localhost-only HTTP API. C++20 / Qt 6.

南方星火 SouthernSpark — Open-source AI agent skills & browser automation tools. Build, share & monetize practical AI workflows. 开源 AI Agent 技能包 & 浏览器自动化工具箱。

Open-source agent harness for knowledge work: long-running AI agents with MCP tools, managed context, subagents and self-improvement in one lean TypeScript-on-Bun binary.

AgentOS — local autonomous agent runtime: plan, execute, verify, review and self-correct engineering tasks (TypeScript core + CLI + Next.js dashboard)

AgentForge: Open-source AI Agent Platform — Build, Deploy, and Orchestrate Intelligent Agents at Scale with Next.js 15, TypeScript & Docker.

As autonomous AI agents are increasingly deployed across enterprise workflows, their tool-calling capabilities (e.g., executing shell commands, sending emails, or modifying databases) expose critical security vulnerabilities. Prompt injections or reasoning loops can lead to catastrophic unauthorized actions—such as dropping database tables or leaki

ASI Agents is a local-first multi-agent desk: Chief plus specialists with skills, approvals, Panic, and a draft-first inbox. On-device AMS Micro 70M routes intents; Hybrid 120M is optional. Escalate to local chat or cloud when needed. Virtual Desktop for apps and browser. Modes: Super · Multi · Pro. Intelligence that works as one on your machine.

Secure, local-first AI development platform with supply-chain security gate. Run coding agents offline (Claude, ChatGPT, Gemini). Desktop IDE, CLI/TUI, MCP plugin, VS Code extension. Nemesis security scanning blocks dangerous code before execution.

Shrimpy is a home-agent framework built on Pi. Keep it Shrimple.

Managed agents on Kubernetes + Self-hosted LLM serving. Own your scalable AI.

Merchant acquiring intelligence suite built on a deterministic agentic workflow for pre-checks, KYB, credit decisioning and unified risk scoring.

A full Web Shell Terminal supporting Web LLM and experimental agentic features

A software factory for AI coding agents. Hand it a spec; it plans, builds, measures the result with checks the agent did not write, feeds the gap back, and stops only when independent checks agree. You stay on the loop, not in it.

Durable Codex-to-Claude delegation with tmux ownership, recovery, collaboration, independent verification, and isolated review lanes.

A personal AI workforce that runs on one Resident. It handles what it can, delegates the rest to Workers, and a System Governor makes the whole thing smarter over time — without trusting self-report.

Claude Code plugin for agentic loop-engineering: PRD design/import, tagged phase planning, and an autonomous TDD execution loop with deterministic escalation and disk-backed state.

KAIROS is a lightweight, self-improving AI agent

ATP (all-the-platforms): one command line for 40+ platforms — social, video, music, chat, shopping, travel and drives. Read, search, post, chat, listen and download, with stable JSON/YAML for scripts and an MCP server for AI agents

Quick menu for Claude Code: every plugin's commands and settings, and Claude Code's own, in one pane and a band above the prompt.

Frontier Agentic Language

A comprehensive batch of lint rules for Go, shared across all your repos.

开源 IP/IPAM 管理系统：机房、机柜、设备、IP 分配、VLAN、交换机、监控告警、客户管理一体化平台，内置 AI 助手（告警解读/NL 查询/RAG/Agentic 巡查诊断）。Flask + React 19 + AntD 6。

👻 A brutally honest AI co-founder that challenges your assumptions and provides contrarian feedback in Claude Code

A simple AI agent that only knows how to run shell commands.

Open-source parametric CAD for humans and AI agents, for Windows and Linux. C++ / Open CASCADE. Windows Studio; Linux headless CLI/MCP.

Agent 档案内容仓 · 34 份档案按 agents / harness / tools 三分区（8 维度固定记录能力边界、运行位置、价格与官方证据；不给总分排名，把「未知」当作合法答案）

Multi-agent orchestrator for software projects — 11 departments, 55 specialist agents. Python, pydantic-ai, OpenCode.

🗂️ Manage tasks effectively with jjtask, leveraging JJ (Jujutsu) for structured planning and execution using DAGs and TODO markers.

Run autonomous AI pentesting agents to find and fix security gaps before attackers do

Open-source software factory, made in Europe. AI agents turn tickets into reviewed pull requests, on infrastructure and models you control.

Pure, plugin-based toolkit for agentic video editing and creation. 124 explicit plugins, driven one at a time by an external agent over CLI, HTTP, MCP or batch.

Agent-first CLI for Proton Mail. Local AI inbox classification, search, and Q&A via Proton Mail Bridge and Ollama. Nothing leaves your machine.

The gold standard for agentic fullstack engineering. TypeScript monorepo template with Claude Code harness, blueprint loops, and immutable quality gates.

MuseAgent: A modular multi-agent music intelligence platform that ingests tracks, extracts spectral features, generates embeddings, tags mood/genre/instrumentation, delivers explainable recommendations with visualizations, exports branded PDF reports, and includes optional enrichment and AI-powered music generation.

Autonomous Maia2 chess agent for the Chessmata platform

A privacy-first, local AI background daemon that autonomously intercepts and patches Python crashes using Ollama and regex.

Ask your data. Get answers your team has verified. Open-source analytics workbench where every answer is labelled Certified, From Metrics or Needs review. Works on your dbt project or straight on your database, with lineage, Apps and an MCP server for Claude, Codex and Cursor.

Clara: a free, private AI assistant that runs on your own PC, with an Android app. Research, browsing, reminders, voice, approvals. Bonsai 2 27B + Hermes agent. 🧪 Beta testers wanted.

Your phone. Your say. An open-source Android AI assistant.

React capabilities for AI assistants, UI agents, and tool calling—with live context, validation, policy, and confirmation.

Sovereign, local-first personal AI desktop OS. Zero-egress multi-agent intelligence powered by Ollama, continuous epistemic decaying memory, proactive Guardian safety firewall, and an Electron HUD.

Horizon: a local-first AI workspace with built-in model loading, tools, artifacts, profiles and agent integrations.

Sci-fi TUI for practicing Go • Gamified AI onboarding assistant

Public Windows downloads and product guide for BIGSPY AI Connect: a local-first AI workspace with project tasks, independent Cute Boss review, model connections and Resource Guardian.

A curated list of desktop AI assistants for macOS, Windows, and Linux in 2026

AI-assisted adaptation guide for iPhone Duo foldable devices. Comprehensive Claude Code skill + docs based on Apple WWDC 2026 Tech Talks.

Personal AI assistant for Discord with memory, Home dashboard, projects, and isolated browser tasks

Self-hosted media suite: raven indexes your video folders and plays them straight from disk, in the browser or its desktop app; luwin is an AI assistant for the friends who share it; rookery signs them in once for all of it. An Nx monorepo over Bun: NestJS, React, Electron, libmpv and Python.

裝在手機上的「對話副駕」：讀取最新訊息、AI 分析意圖並建議回覆，一鍵填入聊天輸入框。｜A conversation co-pilot on your phone: analyzes incoming messages, suggests replies, and fills them into chat apps.

A Postgres workbench in one HTML file: real PostgreSQL (PGlite) in your browser, databases you share as links, and a sub-10 MB binary that connects it to your own Postgres servers. Copy a sample of production to work on locally. Built-in AI (the Clerk) that queries freely but asks before any change.

Piko（皮可）｜个人 Agent 与办公 Agent 工作台。Personal Agent & Office Agent workspace: AI coworkers, group chats, custom identities, skills, proactive updates and multi-agent collaboration. macOS + local web.

电脑上开始，飞书里继续。在 Windows Codex 桌面与手机飞书之间接续同一会话。

AI assistant CLI for managing email, calendars, contacts, and tasks through conversation.

Local-first AI agent workspace with persistent memory, multi-role orchestration & auditable execution. Built for Grok + Codex. 本地优先的多智能体工作台：持久记忆、角色协作、任务可追溯。

🔮 Finally, a tool for your inner life—AI journal, therapist, coach & agent-self that grows with you

Open-source workspace for PostgreSQL data, with an admin UI, access policies, a typed API client, plugins, and live collaboration.

Ruthless automated code review for any tech stack. 31+ frameworks, post-mortem verification, game dev support, anti-AI-slop detection.

An AI-native GitHub experience for modern software collaboration.

A single-file mobile controller for ComfyUI.

Enterprise-grade TypeScript implementation of Google's Agent-to-Agent (A2A) protocol with a bidirectional A2A ↔ MCP bridge, OAuth2/JWT auth, Redis/Postgres persistence, SSE streaming, and OpenTelemetry observability.

Give any AI agent real access to your personal WhatsApp. Search your history, read any conversation, and send messages. One Go binary, read-only by default.

User-set absolute context ceiling for DeepSeek Harness: auto-compacts a session once it reaches a configured token limit. External plugin with a dedicated Settings page.

Read public WeChat Official Account (mp.weixin.qq.com) articles as clean Markdown/JSON. CLI + MCP server + agent skill, plug-and-play via uvx.

A high-performance, zero-copy binary serialization format optimized for SIMD instructions and mmap-readability.

A MCP server that gives AI coding tools accurate, queryable documentation for the AK event-driven MCU framework

Minifying proxy for MCP servers and OpenAPI specs: your agent sees three searchable tools instead of every upstream's full catalog, and overlays fix the tools you don't own.

Install and connect formally verified and executable concepts and rules.

AgentDeck: one local GUI to configure every AI coding agent. Tune behavior, permissions, token usage and cost, skills, subagents, MCP servers and hooks for Claude Code, Codex CLI, Gemini CLI, Cursor, OpenCode and 16 more. Zero dependencies.

Security-focused MCP runtime for reading local repositories and running approved commands in Docker.

MCP connector that verifies a claim's grounding against reality (quotes, citations, code, repos, math) with no LLM in the verification path. Works in Claude and Gemini.

Remote shell + file access for AI agents over authenticated MCP-HTTP — an ssh replacement you talk to over /mcp. Single Rust binary, OAuth 2.1 + Basic auth, auto-backgrounding jobs with paginated logs.

let agents read web sources without a middleman

🧭 AI Agent Skill for lost-item search using Meihua Yishu — works with Claude, Cursor, ChatGPT & Gemini

Control plane for a web chat agent to run constrained, auditable jobs on your own machine — GitHub as durable transport, hard capability boundaries, recoverable runs.

Self-hosted SearXNG metasearch web-search plugin for DeepSeek Harness — no API key, adds a Plugins-page settings card. | DeepSeek Harness 插件：通过自托管 SearXNG 元搜索提供网页搜索，无需 API Key，附带设置页插件配置卡片。

A collaborative hub for AI teams to manage, version, and share configuration assets. Create channels, review changes through pull requests, track activity, and keep your team's prompts, configs, and templates organized in one place. Built with real-time sync and role-based access control.