Agent Team
Run a local multi-agent crew with 44 MCP tools and SQLite-backed coordination from a single npm install when one chat thread is not enough structure.
Overview
Agent Team is a MCP server for the Build phase that runs 13 specialized agents with 44 tools over SQLite via a one-command npm stdio install.
What is this MCP server?
- 13 specialized AI agents collaborating through a shared SQLite coordination layer
- 44 MCP tools exposed via one-command npm install (agent-team-mcp, stdio transport)
- Version 1.5.1; repository AgentTeam on GitHub
- Local stdio MCP—no hosted remote URL in the published server manifest
- Useful when you want persistent multi-role workflows inside Claude Code or Cursor instead of one monolithic assistant
- SQLite collaboration backend
What problem does it solve?
One generalist agent thread makes it hard to keep specialized roles, shared state, and repeatable handoffs when you are building alone.
Who is it for?
Builders who want a structured local agent crew with many MCP tools for complex build-and-review loops.
Skip if: Simple single-tool integrations, fully managed cloud agent fleets, or beginners who only need one HTTP API bridge.
What do I get? / Deliverables
After install, your MCP client can invoke a coordinated agent team with dozens of tools and SQLite-backed collaboration instead of reinventing multi-agent wiring each sprint.
- Running stdio MCP server with 13 agent roles and 44 tools
- SQLite-backed collaboration surface for multi-step agent workflows
Recommended MCP Servers
Journey fit
Canonical shelf is Build agent-tooling because the package is fundamentally an MCP runtime for specialized agents, even though you can reuse it during review and iteration. Agent-tooling fits a stdio MCP server that exposes many tools for role-based agents rather than a single REST bridge or frontend widget.
How it compares
Local multi-agent MCP runtime with 44 tools, not a hosted SaaS MCP generator or a social data API.
Common Questions / FAQ
Who is Agent Team for?
Solo builders and small teams using MCP clients who need multiple specialized agents and shared state without building orchestration from scratch.
When should I use Agent Team?
Use it during Build for agent-tooling workflows, and optionally in Ship review or Operate iterate when you want the same crew to re-run structured tasks.
How do I add Agent Team to my agent?
Install the npm package agent-team-mcp (v1.5.1), add it as a stdio MCP server in Claude Code, Cursor, or Codex, and start the process from your client’s MCP configuration.