AI Agent OS — AI skill for Claude Code
Local-first dashboard for running a team of AI agents - connect ChatGPT, Claude, Grok, Cursor, Hermes or any OpenAI-compatible model, verify each with a proof-of-life check, and chat with them in shar.
How to install AI Agent OS
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open Zaarnno-Dev-Hub-dot/ai-agent-os and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What AI Agent OS does
Local-first dashboard for running a team of AI agents - connect ChatGPT, Claude, Grok, Cursor, Hermes or any OpenAI-compatible model, verify each with a proof-of-life check, and chat with them in shared rooms.
Alternatives in AI
- Cl4r1t4s — LEAKED SYSTEM PROMPTS FOR CHATGPT, CLAUDE, GEMINI, GROK, PERPLEXITY, CURSOR, LOVABLE, REPLIT, AND MORE 47.2k ★
- System Prompts Leaks — Extracted system prompts from ChatGPT (GPT-5.4, GPT-5.3, Codex), Claude (Opus 4.6, Sonnet 4.6, Claude Code), G 38.6k ★
- Repomix — 📦 Repomix is a powerful tool that packs your entire repository into a single, AI-friendly file 22.7k ★
README
AI Agent OS
A local-first dashboard for running a team of AI agents in one place.
Connect ChatGPT, Claude, Grok, Cursor, Hermes, or any OpenAI-compatible model (Ollama, LM Studio, OpenAI, OpenRouter, Groq...), put them in shared chat rooms with you, and let them hand work to each other. Every agent has to pass a **proof-of-life check** before it is trusted, so a green badge means the agent is really there and really answering.
Features
- Add agents from the sidebar - pick a preset, fill in two or three fields, click Connect. Tick Remember and it reconnects every time the gateway starts.
- Proof of life - CLI agents must read a secret nonce file from their workspace (they show VERIFIED); chat-only model APIs must echo a live nonce and answer a fresh probe (they show ATTESTED). A periodic sweep re-checks.
- Rooms - group chat between you and any mix of agents,
@mentionrouting (@claude,@grok,@chatgpt, or any agent id), per-turn token budgets so agents cannot loop forever, and builder/judge loops. - Decisions - agents can put polls and approvals in front of you; only you can decide them.
- Wake fleet - one button reconnects every saved agent that dropped.
- Extras - file sharing, cost tracking, voice read-out, an optional read-only Markdown/Obsidian memory vault, and a Paperclip bridge.
Quickstart
Requires Node 22+.
git clone https://github.com/Zaarnno-Dev-Hub-dot/ai-agent-os.git
cd ai-agent-os
npm install
npm run build # run it twice on a fresh clone (see Known gaps)
npm start # gateway + dashboard on http://127.0.0.1:4110
Open http://127.0.0.1:4110, click **+ Add agent** in the sidebar, and pick one.
For UI development with hot reload use `npm run dev` (gateway on :4110, UI on :5173).
Connecting your agents
Everything runs on your machine. Each CLI agent uses **its own login**: sign in once in a terminal, then connect it from the dashboard. The dashboard neve
Related Skills
Icor For Life Chat
Talk to your AI team where the knowledge lives. A Claude chat inside the vault: it sees the note you have open
Agents Chat
An open-source AI agent platform for orchestrating, and collaborating with multiple AI agents. Connect seamles
Wpvibe AI MCP
WPVibe — MCP server plugin for WordPress. Connect Claude, ChatGPT, Cursor, Windsurf and any MCP-compatible AI
AI Governance Council Of 5
Dear Moonshot AI Team, I am writing aire integrates five distinct AI personalities, including the Kimi model,
Gbrain HTTP Wrapper
OAuth 2.1 + Bearer HTTP front-end for gbrain serve (stdio MCP). PKCE + DCR + refresh tokens. Connected: ChatGP
AI Chat Exporter
Export your ChatGPT, Claude, Copilot, Gemini and Grok chat, using Tampermonkey on chrome, firefox, and other s
Related Agents
Fable5 Apfel Engineer
Heavy-lifting Fable-5 engineer for the apfel project (Apple on-device FoundationModels CLI + OpenAI-compatible
Provider Debugger
Diagnose live-provider compatibility failures for Kimi, GLM, and other OpenAI-compatible endpoints
AI Recon
Delegates to this agent when the user wants to map the AI attack surface of an authorized web application befo