Delego
Require intent-bound authorization, human approval, and a signed audit trail before high-risk agent actions execute.
Overview
Delego is a MCP server for the Ship phase that enforces intent-bound authorization, human approval, and a signed audit trail for AI agent actions.
What is this MCP server?
- Intent-bound action authorization with policy, human approval, and signed audit trail
- PyPI delego v0.2.4 via uvx with delego[mcp] extra and delego-mcp console entry
- DELEGO_HOME scopes policy, signing keys, and ledger (default ~/.delego or project .claude/.delego)
- stdio MCP transport for wiring into Claude Code-style project layouts
- Open-source Delego-Dev/delego repository for policy customization
- Package version 0.2.4 on PyPI registry delego
- Console script delego-mcp with delego[mcp] extra
- Environment variable DELEGO_HOME for policy, keys, and ledger
Community signal: 23 GitHub stars.
What problem does it solve?
Autonomous agents can invoke MCP tools faster than humans can reason about policy, leaving no durable approval or audit record.
Who is it for?
Solo builders graduating to client or team agent workflows who need explicit approval and tamper-evident logging on tool execution.
Skip if: Casual local-only coding with no shared policy requirements or builders unwilling to maintain keys, policy files, and DELEGO_HOME state.
What do I get? / Deliverables
After configuration, risky agent actions flow through Delego policy and optional human approval, with decisions recorded in a signed audit ledger under DELEGO_HOME.
- Configured Delego MCP stdio server with delego-mcp entry point
- Project-scoped policy and signed audit ledger under DELEGO_HOME
- Operational human-approval path for intent-bound agent actions
Recommended MCP Servers
Journey fit
How it compares
Agent authorization and audit MCP server, not a vulnerability scanner or generic secrets manager.
Common Questions / FAQ
Who is Delego for?
It is for developers running AI agents with MCP tools who need policy-bound intents, human approval for sensitive steps, and a signed audit trail.
When should I use Delego?
Use it during Ship security when autonomous tool use could affect production, payments, or shared data and you need accountable gates.
How do I add Delego to my agent?
Run via uvx with delego[mcp] and delego-mcp, set DELEGO_HOME to your policy and ledger directory, add the stdio MCP entry to your client, and restart after policy changes.