Marrow AI MCP server (@getmarrow/mcp). The control layer for AI agents: a named yes before an agent acts, and proof the next agent can reuse. Claude Code, Cursor, Codex, any MCP client.
接入配置
选择客户端，复制配置，再按项目文档补齐环境变量。
claude mcp add marrow-ai-control-layer-for-ai-agents --env MARROW_API_KEY='<MARROW_API_KEY>' -- npx -y @getmarrow/mcp~/.codex/config.toml
[mcp_servers.marrow-ai-control-layer-for-ai-agents]
command = "npx"
args = ["-y", "@getmarrow/mcp"]
env = { MARROW_API_KEY = "<MARROW_API_KEY>" }
# codex mcp add marrow-ai-control-layer-for-ai-agents --env MARROW_API_KEY='<MARROW_API_KEY>' -- npx -y @getmarrow/mcp~/.cursor/mcp.json
{
"mcpServers": {
"marrow-ai-control-layer-for-ai-agents": {
"command": "npx",
"args": [
"-y",
"@getmarrow/mcp"
],
"env": {
"MARROW_API_KEY": "<MARROW_API_KEY>"
}
}
}
}.vscode/mcp.json
{
"servers": {
"marrow-ai-control-layer-for-ai-agents": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@getmarrow/mcp"
],
"env": {
"MARROW_API_KEY": "<MARROW_API_KEY>"
}
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"marrow-ai-control-layer-for-ai-agents": {
"command": "npx",
"args": [
"-y",
"@getmarrow/mcp"
],
"env": {
"MARROW_API_KEY": "<MARROW_API_KEY>"
}
}
}
}claude_desktop_config.json
{
"mcpServers": {
"marrow-ai-control-layer-for-ai-agents": {
"command": "npx",
"args": [
"-y",
"@getmarrow/mcp"
],
"env": {
"MARROW_API_KEY": "<MARROW_API_KEY>"
}
}
}
}.trae/mcp.json
{
"mcpServers": {
"marrow-ai-control-layer-for-ai-agents": {
"command": "npx",
"args": [
"-y",
"@getmarrow/mcp"
],
"env": {
"MARROW_API_KEY": "<MARROW_API_KEY>"
}
}
}
}需要的环境变量
MARROW_API_KEY
把占位值换成你自己的密钥；不要把密钥提交到代码仓库。