MCP Agent
A composable framework for building effective agents using Model Context Protocol that implements Anthropic's Building Effective Agents patterns as reusable workflows. Provides full MCP support including tools, resources, prompts, and notifications, with batteries-included features like structured logging, token accounting, and optional Temporal-backed durable execution for production workloads. Enables developers to connect LLMs to any MCP server through simple, Pythonic decorators and context managers while supporting advanced patterns like map-reduce, orchestrator-workers, and routing.
