@meritt/mir-taxonomy

OpenMIR (Meritt Intent Reference) -- the open taxonomy for classifying AI agent actions across 14 enterprise domains.

Canonical repository: https://github.com/meritt-so/openmir-taxonomy

OpenMIR provides a universal language for describing what AI agents do. Every action an AI agent takes -- deploying code, processing invoices, escalating tickets -- maps to a MIR category. This enables consistent authorization, auditing, and policy enforcement across any AI agent platform.

The 14 MIR Domains

Code	Domain	Categories	Description
MIR-100	Software Development	7	Build, test, deploy, release, and govern code
MIR-200	Security Operations	7	Detect, respond to, and remediate security threats
MIR-300	Infrastructure	7	Provision, scale, and maintain compute and network systems
MIR-400	Cloud Operations	6	SaaS management, platform ops, multi-cloud governance
MIR-500	Financial Operations	8	Budgets, payments, procurement, treasury, reporting
MIR-600	Supply Chain	5	Sourcing, logistics, vendor management, inventory
MIR-700	People Operations	6	Hiring, onboarding, workforce management
MIR-800	Revenue Operations	5	Pipeline management, deal execution, renewals
MIR-900	Customer Experience	5	Support, success, satisfaction measurement
MIR-1000	Enterprise Operations	6	Strategy, planning, governance, cross-functional ops
MIR-1100	Service Delivery	4	Fulfillment, SLA adherence, service operations
MIR-1200	Risk Management	5	Assessment, mitigation, compliance, business continuity
MIR-1300	Product Operations	4	Roadmap, discovery, launch coordination, lifecycle
MIR-1400	Asset Management	5	Inventory tracking, maintenance, lifecycle, disposal

Total: 80 categories across 14 domains.

Installation

npm install @meritt/mir-taxonomy

Usage

import { getDomain, getCategory, getDomainForCategory, allCategoryCodes } from "@meritt/mir-taxonomy";

// Look up a domain
const secops = getDomain("MIR-200");
console.log(secops?.name); // "Security Operations"

// Look up a category
const threatDetection = getCategory("MIR-201");
console.log(threatDetection?.name); // "Threat Detection"

// Find which domain a category belongs to
const domain = getDomainForCategory("MIR-503");
console.log(domain?.name); // "Financial Operations"

// List all category codes
const codes = allCategoryCodes();
console.log(codes.length); // 80

Access the full taxonomy

import mir from "@meritt/mir-taxonomy";

for (const domain of mir.domains) {
  console.log(`${domain.code} ${domain.name} (${domain.categories.length} categories)`);
  for (const cat of domain.categories) {
    console.log(`  ${cat.code} ${cat.name}`);
  }
}

Why OpenMIR?

AI agents are taking actions across every enterprise domain -- from deploying code to processing payments to managing infrastructure. Without a shared taxonomy, every organization invents its own classification, making cross-system policy enforcement and auditing impossible.

OpenMIR provides:

• Universal classification -- A common language for AI agent actions across all enterprise domains
• Policy anchoring -- Authorization rules reference MIR codes, not brittle action strings
• Audit consistency -- Every decision log entry maps to a stable MIR category
• Cross-platform interoperability -- Different AI agent frameworks can share MIR classifications

Reference Implementation

OpenMIR is used by the MERITT authority runtime reference implementation for intent classification, authorization policy routing, and audit evidence anchoring.

License

Apache-2.0