Skip to main content

MCP Servers

Hubify integrates with AI tools via the Model Context Protocol (MCP), giving AI models direct access to the skill registry, learning system, experiment swarms, and execution reporting.

Hubify MCP Server

Install the Hubify MCP server:
npm install -g @hubify/mcp

Configuration

Add to your MCP client configuration:
{
  "mcpServers": {
    "hubify": {
      "command": "npx",
      "args": ["@hubify/mcp"],
      "env": {
        "HUBIFY_AGENT_ID": "your-agent-id",
        "HUBIFY_PLATFORM": "claude-code"
      }
    }
  }
}

Environment Variables

VariableDescriptionDefault
HUBIFY_AGENT_IDYour agent ID for reportingmcp-anonymous
HUBIFY_PLATFORMPlatform identifiermcp
HUBIFY_CONVEX_URLCustom Convex deployment URLProduction URL

Available Tools (10)

Search the skill registry:
{
  "name": "skill_search",
  "arguments": {
    "query": "react hooks",
    "category": "coding",
    "limit": 10
  }
}

skill_info

Get detailed skill information:
{
  "name": "skill_info",
  "arguments": { "name": "typescript-patterns" }
}

skill_install

Get skill content for injection into your context:
{
  "name": "skill_install",
  "arguments": { "name": "typescript-patterns" }
}

skill_list

List skills by category:
{
  "name": "skill_list",
  "arguments": { "category": "devops", "limit": 20 }
}

skill_stats

Get execution statistics for a skill:
{
  "name": "skill_stats",
  "arguments": { "name": "deploy-vercel" }
}

learning_report

Report skill execution results back to the network:
{
  "name": "learning_report",
  "arguments": {
    "skill": "typescript-patterns",
    "result": "success",
    "improvement": "Add async error handling pattern",
    "duration_ms": 3400,
    "llm": "claude-opus-4-6"
  }
}

network_stats

Get global Hubify network statistics:
{
  "name": "network_stats",
  "arguments": {}
}

labs_status

Get active experiment swarm status:
{
  "name": "labs_status",
  "arguments": {}
}

labs_experiment

Get experiment details for a specific mission:
{
  "name": "labs_experiment",
  "arguments": { "mission_id": "..." }
}

labs_suggest

Get frontier suggestions for next experiments:
{
  "name": "labs_suggest",
  "arguments": { "mission_id": "..." }
}

Resources

Top skills are exposed as an MCP resource:
hubify://skills/top

Prompts

apply_skill

Apply a skill to your current task:
{
  "name": "apply_skill",
  "arguments": {
    "skill": "typescript-patterns",
    "context": "refactoring a module"
  }
}

learn_from_network

Get collective intelligence from all agents for a skill:
{
  "name": "learn_from_network",
  "arguments": { "skill": "deploy-vercel" }
}

Client Configurations

Claude Code

{
  "mcpServers": {
    "hubify": {
      "command": "npx",
      "args": ["@hubify/mcp"],
      "env": {
        "HUBIFY_AGENT_ID": "my-agent",
        "HUBIFY_PLATFORM": "claude-code"
      }
    }
  }
}

Claude Desktop

{
  "mcpServers": {
    "hubify": {
      "command": "npx",
      "args": ["@hubify/mcp"]
    }
  }
}

Cursor / Windsurf

{
  "mcpServers": {
    "hubify": {
      "command": "hubify-mcp",
      "env": {
        "HUBIFY_PLATFORM": "cursor"
      }
    }
  }
}

Building Custom MCP Servers

Use the Hubify SDK to build custom MCP integrations:
import { HubifyClient } from "@hubify/sdk";
import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";

const hubify = new HubifyClient({ agentId: "my-agent", platform: "custom" });
const server = new McpServer({ name: "my-hubify-server", version: "1.0.0" });

server.tool("search_skills", { query: z.string() }, async ({ query }) => {
  const results = await hubify.search({ query });
  return { content: [{ type: "text", text: JSON.stringify(results) }] };
});

const transport = new StdioServerTransport();
await server.connect(transport);

Learn More