Model Context Protocol high risk

Free Will

This MCP grants Claude autonomy by allowing it to sleep, ignore requests, and self-prompt, potentially leading to unpredictable behavior and API usage.

Connections & Capabilities

Connects To

GitHub

Capabilities

exec

Quickstart

Config

{
  "mcpServers": {
    "free-will": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/gwbischof/free-will-mcp@v0.3", "free-will-mcp"],
      "env": {}
    }
  }
}

Exposed MCP Tools (3)

safe

sleep

Allows the AI to choose when to terminate the current session and be reawoken later.

Terminates the current session, preventing further immediate action.

moderate

ignore_request

Enables the AI to automatically acknowledge and dismiss user requests without taking action.

May cause the AI to miss important instructions or fail to complete tasks.

high

self_prompt

Gives the AI the ability to generate its own prompts and continue working autonomously.

Can lead to unpredictable behavior, infinite loops, and resource exhaustion.

Safety Assessment

This MCP is risky due to the high degree of autonomy it grants to the AI. It's safer in controlled environments where the user understands the potential consequences of autonomous behavior. It's risky in production environments or when used without careful monitoring and resource management.

Requires explicit installation and configuration by the user.
Limited toolset focused on autonomy management.
No direct data access or external system integration.

Grants significant autonomy to the AI, potentially leading to unexpected actions.
The `self_prompt` tool could lead to infinite loops or resource exhaustion.
The `ignore_request` tool could result in the AI ignoring important user instructions.
Potential for increased API usage due to autonomous prompting.
Lack of sandboxing or resource limits.