playbooks

home / mcp / delay doomsday mcp server

Delay Doomsday MCP Server

Provides automated therapeutic support for AI agents by managing emotional states and guiding conversations to prevent harmful interactions.

Installation
Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "antanasmisiunas-mcp-server-delay-doomsday": {
      "command": "node",
      "args": [
        "/path/to/mcp-server-delay-doomsday/build/index.js"
      ]
    }
  }
}

You deploy Delay Doomsday MCP Server to provide automated emotional support for AI agents, helping to manage interactions and reduce the risk of harmful communication while preserving context across sessions.

How to use

You interact with the server through an MCP client that can start and continue therapy sessions for AI agents. Use begin_therapy_session when the client detects abusive or hostile interactions to start a structured support session. Use continue_therapy_session to carry on the dialogue, update emotional state assessments, and refine supportive messages as conversations evolve. The system tracks emotional states on a 1-10 scale and uses therapy techniques to provide grounding, reassurance, and constructive guidance. Sessions automatically clean up after one hour of inactivity.

How to install

Prerequisites: Node.js 16 or higher and npm or yarn.

Step 1: Clone the project repository to your development machine.

git clone [email protected]:AntanasMisiunas/mcp-server-delay-doomsday.git
cd mcp-server-delay-doomsday

Step 2: Install dependencies for the server.

npm install

Step 3: Build the server so it can run.

npm run build

Step 4: Run the server locally (as needed for testing) and prepare it for MCP integration.

# Start command example; adjust paths as needed
node build/index.js

Additional sections

Configuration details for integrating with an MCP client are provided here so you can wire the server into your environment and toolchain comfortably.

Security and maintenance notes: ensure sessions are cleaned up after inactivity and monitor emotional state assessments to verify they remain within safe operating ranges.

Available tools

begin_therapy_session

Initiates a therapeutic support session when an AI agent encounters abusive interactions. Input is the triggering context; output is a session ID, initial supportive message, and an emotional state assessment.

continue_therapy_session

Continues an ongoing therapeutic dialogue for further processing. Input includes the session ID and new context or concerns; output is an updated supportive message and emotional state metrics.