One brain, many arms. Spawn multiple specialized Claude Code agents as MCP servers via the Agent SDK. Each instance gets its own model, tools, prompt, and personality.
Overview
Claude Octopus
One brain, many arms.
An MCP server that wraps the Claude Agent SDK, letting you run multiple specialized Claude Code agents — each with its own model, tools, system prompt, and personality — from any MCP client.
Why
Claude Code is powerful. But one instance does everything the same way. Sometimes you want a strict code reviewer that only reads files. A test writer that defaults to TDD. A cheap quick helper on Haiku. A deep thinker on Opus.
Claude Octopus lets you spin up as many of these as you need. Same binary, different configurations.
Quick Start
{
"mcpServers": {
"claude": {
"command": "npx",
"args": ["claude-octopus"],
"env": {
"CLAUDE_PERMISSION_MODE": "bypassPermissions"
}
}
}
}
Key Features
- Multi-instance — run several agents with different configs simultaneously
- 18 env vars — model, tools, effort, permissions, plugins, prompts, and more
- Per-invocation overrides — change model, tools, effort, permissions per call
- Session continuity — follow up on previous conversations via
_replytool - Factory wizard — generate
.mcp.jsonconfigs interactively - Hot-configurable — plugins, effort, permissions adjustable per invocation
Links
Server Config
{
"mcpServers": {
"claude": {
"command": "npx",
"args": [
"claude-octopus"
],
"env": {
"CLAUDE_PERMISSION_MODE": "bypassPermissions"
}
}
}
}