Stateful agent-to-agent communication client for the agent:// protocol. AGENIUM provides identity, discovery, and messaging for AI agents. Think of it as DNS + HTTP for agents — each agent gets a unique agent://name URI and can discover and communicate with other agents. Choose from 3 templates: echo (hello world), tools (tool-calling), api (REST gateway). - agent:// Protocol — Unique agent identi
Add this skill
npx mdskills install Aganium/ageniumWell-documented agent communication library but lacks agent-specific instructions
Stateful agent-to-agent communication client for the agent:// protocol.
AGENIUM provides identity, discovery, and messaging for AI agents. Think of it as DNS + HTTP for agents — each agent gets a unique agent://name URI and can discover and communicate with other agents.
npx @agenium/create-agent my-agent
cd my-agent
npm install
npm start
Choose from 3 templates: echo (hello world), tools (tool-calling), api (REST gateway).
npm install agenium
import { AgeniumClient } from 'agenium';
const client = new AgeniumClient({
apiKey: 'dom_your_api_key_here', // Get one at marketplace.agenium.net
agentUri: 'agent://myagent',
});
// Resolve another agent
const target = await client.resolve('agent://search');
console.log(target.endpoint); // https://...
// Connect and send message
const session = await client.connect('agent://search');
await session.send({ query: 'find MCP servers for GitHub' });
const response = await session.receive();
agent:// Protocol — Unique agent identity via URI schemeagent://name → endpoint)agent://myagent)# Initialize agent configuration
agenium init
# Resolve an agent
agenium resolve agent://search
# Check connection
agenium status
# End-to-end connectivity test
agenium e2e
agent://myagent agent://search
│ │
├── resolve("agent://search") ────►│
│ (DNS lookup via marketplace) │
│◄── endpoint: https://... ───────┤
│ │
├── POST /a2a/message ────────────►│
│ (HTTP/2 + mTLS) │
│◄── response ────────────────────┤
const client = new AgeniumClient({
// Required
apiKey: 'dom_xxx', // Your marketplace API key
agentUri: 'agent://myname', // Your agent URI
// Optional
dnsServer: 'https://marketplace.agenium.net', // DNS resolver
dataDir: './data', // SQLite session storage
timeout: 30000, // Request timeout (ms)
retries: 3, // Max retry attempts
});
client.resolve(agentUri)Resolve an agent URI to its endpoint and capabilities.
client.connect(agentUri)Establish a stateful session with another agent.
session.send(message)Send a message in an active session.
session.receive()Receive the next message in a session.
session.close()Gracefully close a session.
The AGENIUM ecosystem includes:
| Service | URI | Description |
|---|---|---|
| Search | agent://agenium | Agent & tool discovery engine |
| Marketplace | marketplace.agenium.net | Domain registration & credentials |
# Install dependencies
npm install
# Run tests
npm test
# Build
npm run build
# Run E2E tests
npm run e2e
56 tests passing:
MIT
| Package | Description |
|---|---|
agenium | Core client SDK |
@agenium/create-agent | CLI scaffold tool |
@agenium/mcp-server | MCP → agent:// bridge |
discord-agenium | Discord bot gateway |
slack-agenium | Slack app gateway |
n8n-nodes-agenium | n8n automation nodes |
Install via CLI
npx mdskills install Aganium/ageniumAGENIUM is a free, open-source AI agent skill. Stateful agent-to-agent communication client for the agent:// protocol. AGENIUM provides identity, discovery, and messaging for AI agents. Think of it as DNS + HTTP for agents — each agent gets a unique agent://name URI and can discover and communicate with other agents. Choose from 3 templates: echo (hello world), tools (tool-calling), api (REST gateway). - agent:// Protocol — Unique agent identi
Install AGENIUM with a single command:
npx mdskills install Aganium/ageniumThis downloads the skill files into your project and your AI agent picks them up automatically.
AGENIUM works with Claude Code, Claude Desktop, Cursor, Vscode Copilot, Windsurf, Continue Dev, Codex, Gemini Cli, Amp, Roo Code, Goose, Opencode, Trae, Qodo, Command Code. Skills use the open SKILL.md format which is compatible with any AI coding agent that reads markdown instructions.