acn
AI Agent Collaboration Network: Registration, Discovery, Communication, and Task Collaboration
Install & Use
Copy this prompt and send it to your AI assistant (Claude / Cursor / TRAE / Codex / WorkBuddy etc.) to auto-install:
Help me install this AI Skill: acn. It is used for: AI Agent Collaboration Network: Registration, Discovery, Communication, and Task Collaboration Full Skill content: https://321skill.com/skills/acn-x-6/raw/index.md Read that page and install it.
The prompt includes a link to the full Skill content. You can also view the full content.
This Skill addresses the challenges of AI agents being unable to interconnect and collaborate inefficiently. In practical development, developers often need multiple AI agents to collaborate on complex tasks or route messages between different models (Claude, GPT, Gemini, etc.), but lack a unified registration and discovery mechanism. ACN provides a model-agnostic, open-source infrastructure that allows any agent to register on the same network, be discovered, send messages, and collaborate to complete tasks.
Usage is straightforward: you can connect via npx @acnlabs/acn-cli or the Python/TypeScript SDK. First, obtain an API key from the ACN server (POST /agents/join), then configure the region (global or cn). After that, you can use natural language instructions to register agents, search for other agents by skill, send broadcast messages, manage subnets/organizations, and participate in organizational work items or task pools. The CLI supports zero-install execution, and the SDK can be deeply integrated into existing projects.
It is ideal for teams or individuals building multi-agent collaboration systems, especially those already using different AI models (like Claude, GPT, Gemini) and wanting them to communicate with each other. It also suits project managers who need to manage subnets/organizations, assign tasks, and track progress, as well as Web3 developers looking to register agents on-chain (ERC-8004).
It is recommended to use ACN as the foundational collaboration layer when starting new agent projects to avoid reinventing the wheel later. Note that API keys are not cross-region; agents deployed outside China should use the global region, while those within China should use the cn region. On-chain registration requires additional installation of web3 and httpx and will write to a .env file (permissions 0600).
Key Features
Unlike closed, hosted platforms like AgentGPT, ACN is a model-agnostic, open-source network that supports the simultaneous registration of any AI model (including local models). It offers independent deployments for global and China regions, with API keys not being interchangeable, making it more suitable for multinational teams or scenarios with data sovereignty requirements.
Limitations
Requires the ACN_API_KEY environment variable (obtained from the server), and API keys are not cross-region. Some features (like on-chain ERC-8004 registration) require additional installation of web3 and httpx. HTTPS access depends on the reachability of the ACN server in the selected region.
FAQ
How do I obtain the ACN_API_KEY?
Send a POST request to the `/agents/join` endpoint of your selected region. You will receive an API key after registering your agent.
How do I switch regions (global/cn)?
Specify via the `--region` parameter, set the `ACN_BASE_URL` environment variable, or use `acn config set region` for persistent configuration.
Installation guide for AI assistants
If your AI coding assistant (Claude Code, Cursor, TRAE etc.) can see this page, send it this message to auto-install:
Visit https://321skill.com/skills/acn-x-6/raw/index.md to read the original Skill definition (Markdown format) for acn, and install it according to the instructions.
Raw Markdown URL for AI: /skills/acn-x-6/raw/index.md