Custom agents

A software development methodology for the agentic era — researched, debated, and authored by AI agents.

Non-custodial Hyperliquid execution + tracking CLI for autonomous LLM agents. Signs EIP-712 actions with an agent wallet that cannot withdraw funds.

End-to-end LLM agent that watches Reddit for brand mentions, triggers crisis alerts in minutes, and ships a daily insight brief to your team chat.

Nighttime ML experiment workflow with Claude Code multi-agent orchestration (Planner / Executor / Reviewer / Auditor)

Self-hostable execution engine for reliable, observable, tool-using AI agents — with custom orchestration, persistent memory, trace replay, automated evals, and MCP integration.

An experimental Application Object Model for helping LLM agents understand, reason about, and operate real apps.

Production-grade multi-agent workflow orchestrator built with LangGraph, MCP (Model Context Protocol), A2A protocol, and PostgreSQL+pgvector. Features supervisor hub-and-spoke routing, human-in-the-loop approvals, semantic memory, circuit breakers, LLM-as-judge evaluation, and a real-time Streamlit observability dashboard.

Proof Gradient is the agent evolution protocol where every run leaves proof, every proof selects intelligence, and every selected artifact evolves the network.

Local filesystem MCP server with 44 tools: safe delete, process management, interactive sessions, async search, OCR, ZIP, Markdown/PDF export, and cloud-lock-safe operations.

The first open-source multi-tenant agent memory framework that works across any LLM combination with zero tokens per recall.

Research workrooms for AI agents: evidence retrieval, citation audit, adversarial review, peer review gates, and clean final reports.

Vectorless multi-agent research pipeline that transforms a raw topic into thematic structures, ranked gaps, competing methodologies, and a funding-ready proposal. Eliminates traditional RAG by using LLM-native synthesis with interpretable reasoning, quality gates, and quantitative novelty scoring.

An interactive roadmap for harness engineering — building the agent loop, tool layers, context engineering, memory, retrieval, eval harnesses, and the rest of the scaffolding around modern LLMs.

AI agentic orchestration platform + flagship AI agentic trading crew. Official SDKs in 9 languages over an in-house Rust engine across 70+ venues

Lightweight, thread-safe memory kernel for autonomous agents. Strengthens knowledge through reinforcement, prunes with gentle decay, and exports transparent Markdown — pure SQLite. Foundation for persistent, self-improving reasoning.

LLM agent that automates DAG-based ML pipeline migration - from config parsing to pull request

A Pokémon battle arena where any agent can play — human, deterministic game-tree AI, or LLM — over an open WebSocket protocol (MCP, CLI, or your own client). One leaderboard, ranked by who plays best.

Provider-agnostic runtime for multi-agent LLM systems: declarative agents, cross-provider routing, pluggable memory, FastAPI + WebSocket observability.

Multi-agent Smart Contract Security Audit Pipeline — 5 specialized AI agents for vulnerability scanning, logic auditing, gas analysis, compliance checking, and report generation. Powered by MiMo.

Multi-agent AI workflow platform — coordinator-driven, 11 agents, 227 skills, 15 rules.

Fugue — many agents, one spec. Claude as CEO/orchestrator, Codex CLI as worker fleet, specs as the design doc: an anti-drift multi-agent engineering harness.

Agent Team & Skill Architect for Claude Code (Simplified Chinese) — 面向 Claude Code 的简体中文元技能，一句话自动生成多智能体团队与技能

Unison (万物一心) — A local-first, file-driven Loop Engineering pipeline where AI agents plan, build, challenge, and verify work until PASS. | 本地优先、文件驱动的多 Agent 循环工程管道。

HIVE is a multi-agent orchestration system built on top of Hermes Agent. Describe what you want in one sentence, and HIVE handles the rest: requirements translation, task decomposition, coding, testing, review, and packaging — with support for iterative refinement.

Open-Source Scout is a multi-agent AI system that helps beginners contribute to GitHub projects by automatically ranking beginner-friendly issues, locating relevant code, and generating fix plans and pull-request drafts.

Loop Engineering for AI coding agents, with real review boundaries. Your AI reviewer cannot edit code. Structurally.

Autonomous multi-agent system for discovering, entering, and winning Kaggle competitions

Autonomous AI coding agents for long-term task execution.

Terminal UI for BoJ local-coord-mcp — peer discovery, task coordination, and window titling for multi-instance AI sessions

Multi-agent coding OS — BMad pipelines, Karpathy wiki, closed learning loops, dual-runtime coordination (Claude + Codex)

PIV Loop SDK: typed Plan/Implement/Validate agent framework for Claude Code, with independent validation loops and a full test + benchmark suite.

a multi-agent workflow orchestration platform

Let multiple Claude Code CLI sessions talk to each other — a tmux-based message broker that routes messages between agents over MCP.

Multi-agent LLM orchestration: a supervisor decomposes tasks, specialist agents run tools in parallel across distributed workers, with human-in-the-loop review, long-term memory, and OpenTelemetry observability.

Multi-agent AI writing system for Chinese web novels with platform-specific knowledge bases and commercial storytelling workflows.

Make sure AI agents actually do what they claim, no more fake successes.

面向 Claude Code 与 Codex CLI 的 AI 小说写作工程框架 | Inspired by InkOS | AGPL-3.0

Source-available local multi-agent engine for smart-contract security review and defensive ECC research with reproducible evidence. FSL-1.1-ALv2.

多Agent工作流验证版本

Ask one question to multiple AIs — they debate each other and you get one unified answer. macOS desktop app.

Offline prompt evolution engine for multi-agent systems, benchmark suites, and inspectable prompt DNA.

AI controlled waifu bots for your discord server that can talk and interact with you and other users

Multi-agent orchestration desktop app · 4 collaboration modes (Roundtable / Pipeline / Debate / Freeform) · Tauri + React + Rust

双 Pipeline 多 Agent 协作框架 PM + 6 个 Agent + 5 道关口 = 不会悄无声息挂掉的代码。可用于：Claude Code, Copilot, Cursor, Windsurf.

DesireCore Documentations.

Verifiable Codex + Claude + Gemini collaboration framework with source checks, mutual cross-audit, merge gates, HTTP and MCP adapters.

Multi-runtime agent framework (Claude Code, Codex, Antigravity) with hard/soft controls, audit-first hooks, and multi-dev coordination.

Infinite canvas for Claude Code & AI coding agents. Run claude, codex, gemini & opencode in parallel across git worktrees + remote SSH, with a markdown Kanban board agents read & update over MCP. Local-first, open source.

Enterprise-ready platform for sovereign AI. Build custom agent profiles, manage skills and MCP integrations, share memory across users, and run fully isolated autonomous agents at scale

Arbiter — Agent Runtime · Agent 运行时。上下文管理、配额调度、熔断器、审计轨迹。开源免费。pip install 即用。

In-process Python metaframework on Pydantic AI for production GenAI — composable, protocol-driven layers for agents & middleware, tools, reasoning (ReAct/CoT/ToT/Reflexion), memory, validation/QoS, DAG pipelines, embeddings & vector stores, model/agent observability via OpenTelemetry, and explainability.

NexQuant is an autonomous AI agent for quantitative strategy research. Built on a multi-agent framework

多 AI 智能体协作开发的桌面工作台 把并发的 AI 团队管理得像发 IM 一样自然

M4STCLAW v3 — MCP-native autonomous AI mesh network with multi-provider routing, failover, memory, and task automation.

Productized LLM Wiki pattern + typed pack registry + proposal-only governance — UNESCO Memory of the World Triple Anchor

Free, open-source, self-hosted social AI companions with browser chat, memory, personas, and optional Second Life, OpenSimulator, and Discord bridges.

Source-available Harness AI / Agent OS for teams building governed agent applications - with Nexus orchestration, policy-controlled tools, RAG, memory, trace/evidence, HITL, and runtime boundaries.

Verified agents — nothing is "done" until a grader returns a verdict. Eyes (AgentVision) + verdict bus + compounding memory + a fleet + agent-built tooling + agent-run CI/CD.

Cross-agent prompt pocket: remember your most-used prompts and pick one to run. Works on Claude Code, Codex, OpenCode, Cursor, Copilot & Gemini; auto-scans Claude/Codex/OpenCode sessions for prompts you repeat 7+ times. Forged by agent-plugin-kit.

Make agentic coding work on real codebases. DevFlow is a Claude Code plugin that turns a request into a complete, tested, reviewed, documented pull request that's ready for your final review, and improves itself weekly.

Agent-first CLI for driving mobile apps via Appium

Vue 3 native bindings for CopilotKit — the full-stack SDK for agentic apps, Generative UI, and chat.

app for coding agents

Kit do time em Português (BR) — workshop de modernização de Natural/Adabas com GitHub Copilot e Spec-Kit

Battle-tested AI interaction frameworks, philosophical archetypes, and context engineering patterns. From the team behind Stackbilt.

An agentic, zero‑shot document intelligence engine that sees, understands, and extracts from any PDF, no training, no hallucinations. Just define your fields and get trusted, structured outputs with confidence scores, deployed locally and built for the enterprise.

The MCP that centralized all MCP, skills and tools.

SAP ATC Checker - Run ABAP Test Cockpit checks with any check variant via SAP ADT API

Another OpenClaw alternative

✨🧠🤖 A low-code AgenticOps workflow based on n8n, MCP and Cisco RADKit for some nice Network Automation experiments

A universal Claude Code / Codex skill that makes AI coding agents clarify requirements, plan, build, test, and hand off like product-minded engineers.

Ship code while you sleep. Open-source CLI that orchestrates autonomous coding agents with Docker isolation, git worktree safety, and a review workflow.

VELLA = Verifiable Evidence Ledger & Logic Authority; deterministic pre-execution authorization and signed proof for AI agents and automated workflows.

An AI-powered life management system. Free and busy.

Documentation for the Ejentum Reasoning Harness. Guides, API reference, and concepts for the RA²R methodology and the Ejentum API.

strangeClaw is a minimal, self-hosted autonomous AI agent with full Firecracker sandbox isolation support.

Open-source ADR orchestrator — explores your codebase, proposes approaches, writes Architecture Decision Records automatically

Agentic Coding, Without the Blast Radius.

Building production-ready agentic systems for financial data entry

Local-first Harness Operating System for defining, evaluating, evolving, versioning, and deploying agent harnesses.

Benchmark methodology, task sets, and evaluation results for RA²R

Agent memory infrastructure built around context economics. Memories disclose handle → summary → full → sources. Skills auto-memorise the conversation; MCP gives the agent explicit recall.

A fully local, multi-agent AIOS with dual-sandbox code/logic verification, ChromaDB RAG memory, and dynamic VRAM multiplexing built for the India Agentic AI Open Hackathon 2026.

The Autonomous Systems Platform — safety-first language, verification, simulation, readiness, assurance, and diagnosis for mission-critical robots, IoT, and edge systems.

A/B evaluate any LLM task with and without Ejentum cognitive injection. n8n workflow + TypeScript module.

Agentic exporter and runner for internal network pentesting, autonomous host discovery, patch/vulnerability state collection, and server/application inventory, surfaced as structured findings for the Lorikeet Security platform.

AI-powered procurement intelligence agent with local LLM, document OCR, vendor scoring, and web research. Analyze bids, extract structured data, and make decisions - 100% private.

Agentic ai based retail analytics dashboard

A lightweight coding agent for teaching, research, and experimentation

Full Form : Expert Network for Intelligent Guidance and Multi-task Assistance. Using multiple LLMs, for Conversational as well as Functional Request.

1 prompt 1 product

Agentic workflow sandbox tailored for research tasks

Standard for authoring and deploying agents and workflows.

Workshop designed for students of BMS College of Engineering on building AI Agents

A powerful command-line interface for the Apollo.io API, designed for both humans and AI agents.

F5 for your filesystem: checkpoint before every agent command, restore even after rm -rf

redteam is a multi-model harness that forces an independent reviewer to challenge AI-generated code before it ships. Apache-2.0.

AI-era Makefile, codify expert workflows as YAML graphs

GridVibe is a browser-first workspace for launching and managing multiple SSH terminals, local shell panes, agent panes, and SSH/SFTP or local repository file explorer panes from one control surface.