io.github.Delego-Dev/delego

Official

by Delego-Dev · Python

Intent-bound action authorization for AI agents: policy, human approval, and a signed audit trail.

Model Context Protocol (MCP) Server: io.github.Delego-Dev/delego

The io.github.Delego-Dev/delego MCP server provides intent-bound action authorization for AI agents. It supports policy enforcement, human approval, and a signed audit trail. The server is positioned for use with AI and LLM agents, addressing confused-deputy risks and authorization workflows.

🛠️ Key Features

Intent-bound action authorization for AI agents
Policy-based control
Human approval step
Signed audit trail

🚀 Use Cases

Authorization for LLM/AI agent actions
Mitigating confused-deputy scenarios
Capturing auditable approval records for agent activity

⚡ Developer Benefits

Aligns agent execution with explicit intent and policy
Provides human-in-the-loop approval
Produces a signed audit trail for authorization events

⚠️ Limitations

Description does not specify supported MCP tools, configuration options, or runtime environment

Topics

ai-agentsauditauthorizationconfused-deputyllm-agentsmcppythonsecurity

Install

Configuration

Environment variables

DELEGO_HOME

Directory holding the policy, signing keys, and audit ledger (defaults to ~/.delego). Set this to keep state project-scoped, e.g. /abs/path/to/project/.claude/.delego.

Arguments

from-flagdefault --from

package-with-mcp-extradefault delego[mcp]

console-scriptdefault delego-mcp

Add to VS Code Add to Cursor

claude_desktop_config.json

json

{
  "mcpServers": {
    "delego": {
      "command": "uvx",
      "args": [
        "delego"
      ],
      "env": {
        "DELEGO_HOME": "<YOUR_DELEGO_HOME>"
      }
    }
  }
}