Nexus MCP
未认领MCP server that lets AI models invoke CLI agents (Codex, Claude, OpenCode) as tools — with parallel execution, retries, and structured output parsing.
01 / 选择客户端
03 / 添加配置
claude mcp add --transport stdio 'nexus-mcp' -- 'uvx' 'nexus-mcp'04 / 在客户端中确认
打开客户端的 MCP 设置，确认服务器已连接并列出工具。复制配置不代表连接成功。
更多AI & ML服务器
浏览完整目录安全概况
已认领和已认证的服务器每周接受一次静态扫描，显示代码能触及的范围（外部服务、环境变量、shell 命令、代理配置目录）以及带有已知公告的依赖。认领此列表即可获得。 安全概况的工作原理
18 个工具中显示 18 个
文档中列出的工具 (18)
内容来自项目文档。服务器握手不会验证每个工具的说明或行为。
agent_backends
List backend capabilities and current availability for one workspace
agent_cancel
Request idempotent cancellation of a queued or active job
agent_continue
Queue another turn on an existing session
agent_diagnose
Queue a sessionless backend diagnostic job
agent_fork
Create a child session when the backend supports forking
agent_list
Page through authorized jobs in one workspace
agent_respond
Resolve a pending approval, permission, question, or form input
agent_result
Read the pending or terminal typed result of one job
agent_review
Queue a typed review operation on an existing session
agent_start
Create a durable session and queue its first turn
agent_status
Read the current normalized status of one job
batch_prompt
Yes
clear_preferences
No
get_model_tiers
No
get_preferences
No
prompt
Yes
set_model_tiers
No
set_preferences
No