lastmile ai Agent MCP Server
Build composable agent patterns and durable workflows using the Model Context Protocol.
Data last scanned yesterday · Reviewed yesterday
Overview
The lastmile ai Agent MCP Server provides a Python framework designed to simplify the construction of AI agents by leveraging the Model Context Protocol (MCP). It handles the lifecycle management of MCP server connections and offers composable patterns that allow developers to implement complex agent architectures based on Anthropic's recommended practices. Beyond basic functionality, the framework supports durable execution through integration with Temporal. This enables agents to pause, resume, and recover from failures, providing a path for scaling simple scripts into robust, high-availability workflows.
Our verdict
With a security score of 44/100, this project is rated as high risk. The security review status is currently unverified, specifically lacking audits for OAuth support, authentication methods, and read-only mode, which are critical for safe deployment. Given that this is a community-maintained project and not an official vendor release, users should exercise caution and perform an internal security review before implementing it in production environments.
- Zero direct dependencies reduces the supply-chain attack surface.
- Implements patterns from Anthropic's agent design documentation.
- Supports durable workflows using Temporal integration.
- Available under the permissive Apache-2.0 license.
- High risk score due to lack of security-focused audits.
- Community-maintained without official vendor support.
- Last commit occurred over six months ago.
Compatibility
| Client | Local | Docker | Remote | Read-only |
|---|---|---|---|---|
| Claude Desktop | ||||
| Cursor | ||||
| VS Code | ||||
| Windsurf | ||||
| ChatGPT |
Frequently asked questions
›What prerequisites are required for the LLM API?
You must add your LLM API key either to the mcp_agent.secrets.yaml file or set it as an environment variable.
›Can I use this framework for durable agent workflows?
Yes, mcp-agent supports durable execution using Temporal, allowing you to pause, resume, and recover agents without changing your API code.
›What MCP features does the mcp-agent support?
The framework provides full MCP support, including tools, resources, prompts, notifications, OAuth, sampling, elicitation, and roots.
Changelog
Badge
Maintain this server? Add the live badge to your README.