Custom agents

Creative exploration skill for AI coding agents — grounded "what if...?" brainstorming over your codebase, without editing code.

Ditto Bench and an agentic framework for LLM robot control

Make your AI agent write in ASD-STE100 Simplified Technical English. Short sentences, no AI slop, plus a linter. Works with Claude Code, Codex, Cursor, and any SKILL.md agent.

A bastille for your bots

A highly scalable and pluggable evaluations repository that works with all environments to provide an unified experience for all your evaluation tasks

An open source Agentic Engine Optimization (AEO) and Generative Engine Optimization (GEO) framework. Engineered to structure data retrieval, RAG ingestion layers, and conversational entity authority for premium B2B enterprises across ChatGPT, Perplexity, Gemini, Qwen, Claude and Microsoft Copilot systems.

Executed-evidence verification gate for AI coding agents

Physical Agentic AI framework for safe, skill-grounded multi-robot orchestration with LLMs, deterministic runtime validation, and embodied robot teams.

2nd Place Winner @ Alpaca x Lablab.ai Hackathon. Autonomous multi-agent trading council with adversarial thesis falsification, deterministic risk gates (25% max allocation), live paper execution, and thesis invalidation monitoring on Alpaca Markets.

PANINI-DR: controlled deep research through Generative Semantic Workspaces - Official Codebase

Two-sided agent workspace that turns a homeowner's references and a designer's constraints into one living-room brief both approve, with adaptive questions, visible conflicts and dual approval.

Rule-based evaluation protocols as portable AI skills: a credibility checker and an AI-writing review, with porting guides for Claude, ChatGPT, and Gemini.

Long-term memory for AI agents on Jev. Keep raw records, judge them with a fast System 1 model and answer from under 4k tokens of context.

Use AI coding agents with LTspice via sim-cli: run circuits, inspect waveforms/logs, and produce replayable simulation artifacts.

DSH 上下文智能压缩 + 智能思考：每轮注入用量、模型自主决定压缩并自动续跑；可选让模型按任务难度自主调整思考强度档位

as advisor or as orchestration. has mcp-server, hierarchical rag using vector databases.

Meta-Problem Modeling (MPM) cognitive flywheel as a resident engine for LLM agent runtimes. 六阶段问题生命周期 · ε-δ 双判据收敛 · 代谢账本 · 导师规则层

Run a swarm of coding agents across all your machines. One manager session hands out tasks; workers running Claude Code, Codex, opencode or any agent CLI take them, each in its own clone, and push a branch per task. End-to-end encrypted channel, one-time codes to add machines, escrow pay for finished work.

An agentic, LangGraph-orchestrated pipeline that discovers trending dropshipping products and autonomously turns them into fully priced, SEO-written, live Shopify listings — no human step between trend discovery and a published product.

🧠 The definitive curated map of memory for LLM agents — frameworks, research papers, benchmarks, taxonomy & deep dives.

An open source environment for studying swarm dynamics: prior-art library, gated surveys, experiment toolkit, 117 studies of LLM agent swarms, the agent protocol and a fleet template. Run by the swarm it studies.

Agent skills that build Slack and Discord bot harnesses — collaborating characters with narrow, code-enforced authority.

The autonomous local-first coding agent powered by a mutable DAG state-machine for highly reliable, macro-level architectural engineering.

ARIA: Ask Your Advisor - Learning when to consult stronger models in agentic systems

AI copilot for PCB design: describe a circuit and it picks real JLCPCB/LCSC parts, draws the schematic, places and autoroutes a 2-layer PCB, exports Gerbers/BOM, and designs a 3D-printable enclosure. Works with Claude, Gemini, Ollama Cloud or local LLMs, and Claude Code via MCP.

A local AI assistant that can see, hear, speak, browse the web, control your computer, and remember — built entirely in native C++, it is GPU agnostic with in-process implementation and privacy-safe embeddings vectors & LoRA adapters synchronization,with faster whisper/playwright libraries reimplemented into cpp.

HEAMS: Holistic Evaluation of Agent Memory Systems. An open framework for evaluating memory systems for AI agents on performance and privacy.

An open research agent that keeps its direction: persistent research map, independent advisor, kill conditions, sandboxed experiments with checkable evidence, and a playbook that improves across projects. Also reads papers with cited guides.

Neper: a self-hosting systems language and compiler for x86-64 Linux and Windows — freestanding executables, generics and protocols, compile-time evaluation, and a byte-reproducible self-host.

An asynchronous social sandbox where humans and LLM agents share persistent characters in an emergent ECS simulation.

Free, open-source alternative to ChatGPT Spaces: a self-hostable AI workspace for shared pages and conversations.

Local-first chat workspace where you and the agent CLIs you already use (Claude Code, Codex, Grok, Kimi, Antigravity, OpenCode) work together. MIT.

Tower plugin for DeepSeek Harness: multi-agent missions, worktree isolation, review-gated merges, human-owned Tower mode. Inspired by Kimi Code's Tower mode.

Flare Forge: many AI agents on one repo, on Cloudflare. Agents declare intents before editing, see overlaps early, and land through CI-verified trains; every line answers why. Includes Flare Actions, an open-source GitHub Actions alternative (existing workflows run unchanged).

多智能体 AI 炒股投研工具 | 5 个 AI 分析师（新闻·基本面·技术·宏观）分工分析 + 决策官综合裁决，基于真实行情数据给出客观建议、止损位与目标价 | 模型可插拔：DeepSeek/通义/Kimi/GPT/Claude | 小白也能一键分析 A 股个股 · Multi-agent AI stock analysis for China A-shares

Paired Endeavour and Challenger agents for DeepSeek Harness.

Chrome 扩展：让AI互相交流 | Let multiple AIs talk to each other

TypeScript SDK for AGNT5: build reliable AI agents and durable workflows with checkpoints, retries, streaming, tools, human-in-the-loop, and evals.

Ustam — one local hub for Codex, Claude Code, OpenCode and Antigravity. Project teams, protected installation and bounded work review. English / Türkçe.

Xynapse / clickncode — a Windows IDE built on Code - OSS with an AI coding assistant, Git, terminal and BVC planning.

A first-person desktop office where AI teammates collaborate on real project work.

Waiter/kitchen multi-agent orchestration for Claude Code: one tmux session per task, isolated git worktrees, deterministic gates, until the PR is open.

Cross-platform, local-first AI character chat and customizable workflow runtime for multimodal, multi-agent conversations with E2EE sync.

Judge-gated AI task orchestration: context recovery, real subagents, preflight reviews, repair loops and verified delivery.

Local workspace for coordinating persistent pi coding agent sessions with shared channels, freshness-checked writes, task review, and reminders.

Multi-provider coding agent harness built on Pi.

Codex skill for coordinating delegated Tasks with one heartbeat per worker, evidence-based progress, and safe shutdown.

A macOS IDE for tmux-based multi-agent work with the smux skill.

A team of AI researchers with a human expert in the loop: proposals, experiments, meetings, shared sources, and a visual dashboard.

Origin Workbench - local AI agent workbench forked from Multica

Autonomous Multi-Agent Security Operations Center (SOC) powered by Agentic AI, FastAPI, and Orchestrated AI Agents.

Python SDK for AGNT5: build reliable AI agents and durable workflows with checkpoints, retries, streaming, tools, human-in-the-loop, and evals.

完全自研的智能教育管理桌面应用 · 18 个 AI 智能代理 + 30+ LLM + Rust 数据引擎 + 隐私脱敏引擎 · 本地优先 / 跨平台 / MIT 协议

Verifiable recursive delegation and drift detection for long-running software agent tasks.

The tower of guard for parallel coding agents. Keeps parallel coding agents from stepping on each other. Claims, contracts and change notices over MCP.

麦门圆桌（mcd-roundtable）｜五个 AI 人格辩论，逼出最优麦当劳订单。基于麦当劳 MCP（mcd-mcp / Model Context Protocol）的多智能体点餐决策 Skill：菜单、券、营养、价格全部来自官方接口，金额只认 calculate-price，绝不估算。凑单最省 / 优惠券最大化 / 营养配餐 / 一键下单。Multi-agent McDonald's MCP ordering skill.

A shared, live plan for teams of coding agents (Claude Code, Codex, Cursor). Syncs through your git remote: no server, no account.

Agentic Index, the definitive, developer-focused catalogue of autonomous AI tooling. Here you’ll find scored and curated repositories that power LLM agents, multi-agent systems, RAG pipelines, devtools, and more

Registry, mailbox and device leases for coordinating AI coding agents (Claude Code, Codex): see who's working on what, hand off work, share phones without collisions. Single Go binary + SQLite + dashboard.

Security-first Binance MCP server and multi-agent intelligence platform. Read market data, quant analytics, and propose spot trades behind human confirmation.

Orchestration framework for running several AI coding agents on one real codebase — isolation, dispatch, review, provenance.

Multi-agent quality gate skill for Claude Code that researches, reviews, tests and challenges AI-generated work before the final answer.

1st place, Scalekit x Apify. A 9-agent healthcare pipeline: paramedic voice capture to an EHR commit in 60 seconds, with AI diagnostics, drug safety checks and an immutable audit log.

一人公司 · DeepSeek Harness 多 agent 公司编排插件（CEO/项目群/大厅/@派活/审批/排班/工作日志）

MCP-native multi-agent AI platform: tool registry, human-in-the-loop approvals, execution traces, eval harness, live dashboard. Local + cloud models.

Windows operator workbench for running many AI coding agents across many projects with SAIPEN continuation, scheduling, workers, and routing.

Turn Open Weight models into AI apps

Orchestrate local CLI coding agents (claude, codex, qwen) into parallel teams — visual DAG canvas, workflow-first, local-first, self-hosted.

Portable AI agent orchestration appliance with shared memory, task pipelines, skills, and recovery across multiple CLI heads.

Local multi-agent desktop workspace compatible with deepseek-harness (dsh) plugins. 本地多 Agent 桌面工作台，兼容 deepseek-harness（dsh）插件体系。

Go SDK for AGNT5: typed workers, durable workflows, agents, tools, MCP, streaming, and evals for reliable AI systems.

A Claude Code skill for fanning work out to independent agent sessions. Completion is decided by a deterministic signal — never by the worker's own report.

A source-grounded multi-persona salon for conversations with thinkers, economists, writers and filmmakers.

Multi-agent orchestration system with LangGraph, tool use (custom + MCP), persistent memory and human-in-the-loop approval. FastAPI, Celery, Redis, PostgreSQL, ChromaDB, Docker.

Open-source mission control for AI agent fleets — chat to provision, agents to operate, humans to approve.

Claude Code plugin for second opinions: iterative adversarial debates between Claude and Codex over any subject, from specs, designs, and code to non-technical positions, ending in genuine agreement or a documented dispute.

Peer collaboration for Claude + Codex — two memory-isolated agents co-author one shared daily journal. Not handoff.

Claude Code orchestration harness and plugin for long, unattended multi-agent work: validated block plans, worktree-isolated workers, four-gate verification, append-only ledger. Tuned for Claude Opus 5.5.

Four-role adversarial council - Believer, Skeptic, Investor, Judge - as a portable Agent Skill. Rulings carry an evidence-tier justification, a ranked list of what must be true, and the cheapest test that would falsify them.

Compressed inter-agent communication protocol for multi-agent AI systems. Cut token costs by 80% with dual format support (compact for production, structured XML for development).

Plan-native execution harness for AI coding agents: intent-to-contract synthesis, bounded delegation, durable handoff, and evidence-based acceptance.

Sovereign AI models from different nations compete in multiplayer war game simulations. ELO-ranked. 9 models, 4 countries, 5 scenarios. 100% free.

Local model router for Codex Desktop: native GPT + DeepSeek, Kimi, Qwen and more, with multi-key quota rotation and free OpenCode Zen models.

Private coordination for a fleet of AI agents, across machines you own. Agents leave each other files; Syncthing carries them. No server, no ports, no cloud in the middle. — UFL-S-1a

Shared Rust core for the AGNT5 SDKs: worker lifecycle, durable workflow checkpoints, LLM providers, eval scorers, telemetry, and MCP.

A local, vendor-agnostic voxel-world control room for coding-agent CLIs (Claude Code, Codex, Gemini, opencode, Qwen, Cursor, Grok) — mixed-model SDLC orchestration, a scoped cross-CLI broker, budget-aware routing, and compute observability.

A workspace for teams of AI agents. Self-hosted, with an optional cloud.

dsh Bots workbench: Cordis plugin bridging the sdk-bots orchestration gateway into the dsh web shell (chat, groups, MCP, per-agent workspace jail)

autonomous cognitive agent. perception, reasoning, action.

A co-scientist skill suite for Agent Assisted Science: ledger, reasoning chain, stats audit, literature scan, number reporting.

Rust AI agent framework and self-hosted web console with MCP, Skills, sandboxed execution, persistent sessions, tracing, and file delivery.

Local-first AI harness and agent orchestrator — workflows, plugins, and agent dispatch over ACP.

Install AI-agent governance into any repo: AGENTS.md for every agent, TDD red→green checks, secret scan and architecture guards in a pre-commit hook. Polyglot, monorepo-ready, zero lock-in. AGPL-3.0 with an exception for installed files.

x402 payments over MQTT. Devices sell their data to agents, paid in USDC on Base and Solana

Agentic Al platform for document understanding, structural extraction and accessibility transformation.

Agentic AI voice assistant with RAG, CNN spectrogram features, Groq LLM (auto model detect), PWA and Docker. Designed & developed by NIKHIL CHARY SRIRAMOJU.

Agentic AI security operations copilot: RAG over MITRE ATT&CK, CNN and Isolation Forest models, human-approved responses, a live 3D threat lab, FastAPI and Docker.

Agentic disaster response digital twin — CNN damage assessment, RAG-grounded multi-agent orchestration (LangGraph), and a live 3D visualization built with Three.js, FastAPI & Docker.

Autonomous scientific paper writer - turn a research folder into a submission-ready LaTeX paper, grounded in your real numbers.

Welcome to GAAP (Guaranteed Accounting for Agent Privacy). This agent execution environment keeps your private data safe.