Claude Code

Provides orchestration of Claude Code agents with ADDER+ task workflows across isolated sessions.

python

1

GitHub Stars

python

Language

5 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "nexus-digital-automations-claude_code_mcp_2": {
      "command": "/ABSOLUTE/PATH/TO/Claude_Code_MCP/.venv/bin/python",
      "args": [
        "src/main.py"
      ],
      "env": {
        "PYTHONPATH": "/ABSOLUTE/PATH/TO/Claude_Code_MCP/src"
      }
    }
  }
}

You deploy and manage Claude Code MCP to orchestrate multiple agents across isolated processes, with session-bound workspaces and secure inter-agent communication. This MCP server enables autonomous agent collaboration on complex codebases while maintaining strong security, audit logging, and persistent state across restarts.

How to use

Use an MCP client to connect to the Claude Code MCP server and manage agents, sessions, and messages. Create agent instances to perform tasks, spawn sessions tied to specific codebases, and send messages to agents with ADDER+ prepending for consistent task coordination. Monitor session status to see all agents in a session and use the provided tools to manage conversations and workflows.

Typical workflows include creating agents with specialized system prompts, assigning tasks via instructions and TODO.md files, and coordinating actions through inter-agent messages. Each agent runs in its own process to guarantee strict isolation, and every interaction is logged for security auditing. Your operations are controlled by a role-based permission model to restrict access to sensitive actions.

How to install

Prerequisites: you need Python installed and an MCP client capable of loading STDIO servers (such as uv). Ensure Python is available in your environment path.

Step-by-step setup complete commands you can run to start the MCP server locally.

Install dependencies using the MCP client workflow

uv sync

Start the MCP server

python src/main.py

Configuration and usage notes

You configure the MCP server as an STDIO server in your MCP client by providing the complete command that launches the server.

Available tools

create_agent

Create a new Claude Code agent instance with a unique identity and system prompts.

delete_agent

Remove an existing Claude Code agent from the system, reclaiming resources.

create_session

Create a new session tied to a specific codebase root path for agent orchestration.

get_session_status

Return the status of all agents within a given session to monitor health and progress.

delete_session

Remove an entire session and all agents associated with it, cleaning up state.

send_message_to_agent

Send a message to a targeted agent with ADDER+ prepending for task coordination.

clear_agent_conversation

Close the current iTerm tab associated with an agent to reset context.

start_new_agent_conversation

Open a new iTerm tab to begin a fresh conversation with an agent.