MultiAgentMCP
Related Servers
Alternatives to MultiAgentMCP
No user-submitted related servers found.
Related Servers
- AlicenseAqualityDmaintenanceA local MCP server that lets Claude Code orchestrate Codex CLI and OpenCode CLI as sub-agents with strict security, concurrency, and output limits.41MIT
- AlicenseAqualityFmaintenanceMCP server for running external coding agents as background tasks inside Claude Code. Supports multiple backends including Codex, Grok, GLM, DeepSeek, and more.7MIT
- AlicenseNot gradedqualityBmaintenanceA stdio MCP server that lets any coding agent call any other as a subagent, exposing ask_ and dispatch_ tools for Claude Code, Codex, OpenCode, Antigravity, pi and HTTP endpoints. It enforces recursion depth guards, non-escalating permission ceilings and HMAC-attested results, and can hand off to a fallback agent when a rate limit is hit.132 npm3MIT
- FlicenseNot gradedqualityCmaintenanceA local MCP server that coordinates Claude Code CLI as subagents, providing tools to spawn, message, wait for, and interrupt agents.1-
- AlicenseAqualityCmaintenanceA local stdio MCP service that unifies coding agents like Codex and Claude into cs_agent_* tools, enabling the root agent to create, invoke, and manage child agents with recursive delegation.1456 npm2MIT
- AlicenseAqualityDmaintenanceMCP server that lets Claude Code drive the local Codex CLI as a sub-agent for concurrent queries and optional file/shell actions, using the CLI's existing login and sessions.41MIT
TDQS
Scored across 14 tools
Most tools have distinct roles, and descriptions explicitly separate the launch trio (run_agent = blocking single, start_agent = background preferred, run_agents_parallel = fan-out). However, get_run vs get_agent vs list_agents overlap conceptually (execution record vs job state vs job listing), and run_agent is fully reproducible via start_agent + wait_agents, which invites misselection.
Nearly everything follows verb_noun snake_case (list_*, get_*, run_*, start_*, cancel_agent, close_conversation, send_agent_message). Minor deviations: usage_summary lacks a verb, run_agents_parallel adds a modifier, and singular/plural is inconsistent across get_agent vs wait_agents/list_agents.
14 tools sits squarely in the well-scoped range, and each tool maps to a real orchestration step (discovery, launch, wait, inspect, cancel, cleanup, observability, policy). No tool appears redundant enough to drop.
The agent lifecycle is well covered end to end: discover providers/models, start/run, wait, fetch results, message, cancel, close, plus usage and policy introspection. Gaps: policy is read-only (get_policy with no set/update), and there is no way to retrieve full conversation history/transcripts before close_conversation discards them.