emilia-protocol

Authority control plane for autonomous work. EMILIA Gate enforces finite customer-owned mandates at protected executor boundaries; the open protocol keeps evide

What is emilia-protocol?

Authority control plane for autonomous work. EMILIA Gate enforces finite customer-owned mandates at protected executor boundaries; the open protocol keeps evidence verifiable.

Overview of emilia-protocol

What it is

EMILIA Protocol is an authority control plane for autonomous work. It provides EMILIA Gate, which enforces finite customer-owned mandates at protected executor boundaries. The open protocol keeps evidence verifiable. It is designed to ensure that an agent's actions are authorized before a protected tool runs, with independently checkable evidence under the verifier's own trusted keys and rules.

Key features

  • EMILIA Gate checks authority before a protected tool runs, preventing unauthorized actions.
  • The protocol produces independently verifiable evidence using the verifier's own trusted keys and rules.
  • Supports integration with existing identity providers, agent frameworks, and business systems.
  • Provides MCP and HTTP interfaces for protecting tools, as well as wrappers for Hugging Face smolagents and GitHub Merge Gate.
  • Includes a security case with 35 security claims over 264 hashed evidence files, 20 Tamarin lemmas, and 10,700+ automated tests across 650+ files.
  • Separates identity, policy, and authority into distinct layers: credentials grant reach, policy defines general allowance, authority defines the exact job for an autonomous worker.
  • EP Core exposes three interoperable objects: Trust Receipt (attributable evidence), Trust Profile (structured trust state), and Trust Decision (policy-evaluated result).

Use cases

  • Protecting a supplier payment approval: an agent prepares a payment, someone approves it, but if the amount or destination changes, the earlier approval must not release the changed payment. EMILIA Gate binds approval to specific details and refuses changed or replayed requests.
  • Protecting any tool that holds a real provider credential: the owner defines allowed work and when fresh approval is required, allowing agents to work within limits without human approval for every call.
  • Consequence admission conformance testing: the AEB-1 pack tests the evidence-to-effect boundary at the last control point before a consequential action.

Getting started

With Node.js 20.19 or newer and npm installed: bash git clone https://github.com/emiliaprotocol/emilia-protocol.git cd emilia-protocol npm ci FAST=1 node examples/mcp/payment-server.mjs This runs a local demonstration with generated signing keys, in-memory consumption, and a mock payment tool. It does not perform a real human ceremony or move money.

License and pricing

Licensed under Apache-2.0. Open source and free to use.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/emiliaprotocolemilia-protocol.svg)](https://yunodb.com/tools/emiliaprotocolemilia-protocol?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/emiliaprotocolemilia-protocol?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/emiliaprotocolemilia-protocol.svg" alt="Featured on YunoDB" height="20"></a>