kailash-mcp

Implements MCP servers with stdio, SSE, and HTTP transports for AI agent tool integration.

Updated Apr 2, 2026
One-click install
npx skills add https://github.com/Dchuuuuuu/disease-risk-classifier --skill kailash-mcp-dchuuuuuu
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: kailash-mcp
Source: https://github.com/Dchuuuuuu/disease-risk-classifier/tree/main/.claude/skills/05-kailash-mcp
Command: npx skills add https://github.com/Dchuuuuuu/disease-risk-classifier --skill kailash-mcp-dchuuuuuu

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Building Model Context Protocol servers that expose workflows, tools, and data resources to AI agents requires handling transports, authentication, schema validation, and testing, which is error-prone without structured guidance. ## Core Features & Use Cases - MCP Server Implementation: Create servers with structured tools, resources, and prompts using the Kailash Core SDK. - Transport Configuration: Set up stdio, SSE, HTTP, and WebSocket transports with retry and reconnection logic. - Authentication & Security: Apply API key, JWT, OAuth 2.1, and HMAC authentication patterns for production servers. - Use Case: Expose a Kailash workflow as an MCP tool so Claude Desktop or another AI client can call it with validated inputs and progress reporting. ## Quick Start Ask the AI to create an MCP server that exposes a Kailash workflow as a structured tool using the stdio transport.

Frequently Asked Questions about kailash-mcp

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create an MCP server in Python with Kailash?▼

Import MCPServer from kailash.mcp_server, instantiate it with a name, and register tools using the @server.tool decorator with type-hinted functions. Call server.run() to start it with the stdio transport by default.

Which MCP transport should I use: stdio, HTTP, or WebSocket?▼

Use stdio for local development and CLI tools since it is simplest and lowest latency. Use HTTP for production deployments and load-balanced services, and WebSocket for real-time streaming or progress updates.

How do I add authentication to an MCP server?▼

MCP authentication supports API keys in headers, Bearer tokens, JWT with claims, and OAuth 2.1 client credentials. Store credentials in environment variables, use HTTPS in production, and rotate keys regularly.

Does Kailash MCP support tool input validation?▼

Yes, structured tools accept JSON Schema definitions for inputs and outputs, including enums, nested objects, conditional schemas, and format validators. Validation runs automatically before tool execution and reports errors to the agent.

How do I test MCP servers without mocking the protocol?▼

Use a three-tier strategy: mock provider only for Tier 1 structure tests, real MCP servers for Tier 2 integration tests, and production-like multi-server setups for Tier 3 end-to-end tests. Since v0.6.6, real MCP execution is the default.