Agent Rack
未认领MCP server that runs CLI AI coding agents (Claude Code, Codex, opencode, Antigravity) as sub-agents from any MCP client — background sessions, terminal dashboard, structured code review
安装
npx agent-rack setup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
32 个工具中显示 32 个
文档中列出的工具 (32)
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_list_available
No parameters. Lists every configured agent and whether its binary is on $PATH.
agent_run
agent_session_cancel
agent_session_create
Same execution parameters as agentrun (agent, prompt required; workspace, mode,
agent_session_delete
agent_session_logs
agent_session_send
agent_session_status
agent-rack-review-handling
An agentreview call returns
agent-rack-tool-selection
Delegating any task to a sub-agent through agent-rack
agents
claude
claude mcp add agent-rack -- node <resolved-bin-path> start. --scope maps directly to Claude Code's own -s local\
codex
codex mcp add agent-rack -- node <resolved-bin-path> start (global only — codex has no project-scope flag).
config-check
cursor
Merges an mcpServers.agent-rack entry into Cursor's mcp.json, plus copies agent-rack's two guidance skills into Cursor's skills/ directory. --scope user (default) writes to ~/.cursor/; --scope project writes to <project>/.cursor/ instead.
danger-full-access
--sandbox danger-full-access + escape hatch
desktop
Merges an mcpServers.agent-rack entry into Claude Desktop's config (macOS only).
followUpMode
Agents
formatResponse
runs once the process exits, reducing all accumulated events into a
getCLIArgs
builds the argv for a single run, given the prompt and an optional mode.
line_start
/lineend may be 0 for whole-file, deleted-file, or architectural findings.
live
opencode
opencode
Merges an mcp.agent-rack entry into opencode's config ($OPENCODECONFIGDIR, else $XDGCONFIGHOME/opencode, else ~/.config/opencode) — note this target uses a different config shape ({ type: "local", command: [...] }) than the others. Global only.
OpenCode
agent-rack install --target opencode
opencode_run
identical to agentrun but with agent pre-filled.
parseChunk
is called on every stdout/stderr chunk as it streams in; return zero or more
read-only
--sandbox read-only (enforced)
resume
claude, codex