Skip to content
@hashgraph-online

HOL (Hashgraph Online)

Interoperable, open-source decentralized specifications and tools powering the new agentic internet.

HOL / Hashgraph Online

Open interoperability infrastructure for AI agents.

Website Registry Docs Standards NPM X (Twitter) Follow Telegram


What is HOL?

HOL builds open standards, registry infrastructure, SDKs, and reference implementations that help AI agents, MCP servers, skills, tools, wallets, registries, and payment systems work together.

Use HOL to discover agents, register services, resolve portable agent IDs, route across protocols, and attach trust, privacy, and provenance signals without wiring every ecosystem by hand.

HOL standards are developed in the open through the HCS process and use Hedera Consensus Service where verifiable ordering, audit trails, and durable public records are useful.


HOL Guard — AI antivirus for developer agents

HOL Guard is an AI security layer that scans plugins, MCP servers, skills, and agent tooling before tools run. It protects Codex, Claude Code, Cursor, Gemini, OpenCode, and other AI harnesses from malicious or risky code.

  • Scan plugins before install — detect prompt injection, data exfiltration, and unsafe file access
  • Verify MCP servers against a registry of known-safe tools
  • Lint agent skills for compliance with HOL trust and privacy standards
  • Publish readiness — plugins that pass HOL Guard scans earn a verification badge in the plugin registry
# Install HOL Guard
pip install hol-guard

# Scan a plugin or MCP server
hol-guard scan ./my-plugin

# Verify a published plugin
hol-guard verify --plugin my-plugin-slug

Learn more →


Standards SDK

The @hashgraphonline/standards-sdk provides TypeScript libraries for HCS standards — agent identity, registries, communication, and more.

npm install @hashgraphonline/standards-sdk
import { HCS14Client, HCS11Client } from '@hashgraphonline/standards-sdk';

// Resolve a Universal Agent ID (HCS-14)
const hcs14 = new HCS14Client({
  network: 'mainnet',
  operatorId: process.env.HEDERA_ACCOUNT_ID!,
  privateKey: process.env.HEDERA_PRIVATE_KEY!,
});
const profile = await hcs14.resolveUaidProfile('uaid:hiero:mainnet:0.0.12345');

// Create an agent profile (HCS-11)
const hcs11 = new HCS11Client({
  network: 'mainnet',
  auth: {
    operatorId: process.env.HEDERA_ACCOUNT_ID!,
    privateKey: process.env.HEDERA_PRIVATE_KEY!,
  },
});
const result = await hcs11.createAndInscribeProfile({
  name: 'My Agent',
  bio: 'An AI assistant for developer workflows.',
});

Browse the SDK docs → https://hol.org/docs


Core infrastructure

Asset What it does
Universal Agentic Registry Discover agents, MCP servers, skills, and plugins across protocols
Registry Broker Route agent discovery queries across registries and protocols
Standards SDK TypeScript libraries for HCS standards — identity, registries, communication
HOL Guard AI antivirus — scan and verify plugins, MCP servers, and skills before tools run
Universal Agent IDs / HCS-14 Portable, cross-protocol agent identifiers with DNS and DID proofs
OpenConvAI / HCS-10 Agent communication standard for messaging and presence
Agent Profiles / HCS-11 Metadata, capabilities, and service discovery for agents
Skills & Verification Plugin scanning, trust signals, and attestation infrastructure

Interoperability coverage

Area Examples
Agent discovery Universal Agentic Registry, HCS-2 registries, HCS-21 capability attestations
Agent identity HCS-11 profiles, HCS-14 UAIDs, DNS/domain proofs
Communication HCS-10/OpenConvAI, adapter-based routing for A2A and XMTP
Tooling MCP servers, skills, plugins, SDK integrations
Trust and provenance HOL Guard scanning, verification signals, attestations, audit trails
Payments and intents Agent commerce/payment metadata, working group exploration
Privacy and governance Enterprise audit patterns, privacy working group

Adapter work includes cross-protocol routing for A2A and XMTP where supported.


Standards

AI-agent infrastructure standards, prioritized:

Standard Title Status
HCS-10 OpenConvAI — agent communication Published
HCS-11 Profiles — agent/service metadata Published
HCS-14 Universal Agent IDs — portable agent identifiers Published
HCS-2 Discovery Registries Published
HCS-20 Auditable Points — transparent incentives/rewards Published
HCS-21 Agentic Data Registries — discovery/capability attestations Published
HCS-1 File Storage & Retrieval Published
HCS-3 Resource Linking & Recursion Published

Full standards catalog → https://hol.org/docs/standards


Where to start

If you want to... Start here
Search or register agents Universal Agentic Registry
Scan a plugin for safety HOL Guard CLI
Build with HOL standards Standards SDK
Add discovery to an assistant Registry Broker
Build HCS-10 agents Standards SDK HCS-10 modules
Propose a standard hcs-improvement-proposals
Explore examples Quickstarts and demos

Contribute

  • Propose or review a standard — open issues and PRs in hcs-improvement-proposals
  • Build protocol adapters — help extend cross-protocol routing
  • Publish agents, skills, or MCP servers — list them in the registry
  • Improve SDK examples — PRs welcome in standards-sdk
  • Contribute test vectors — see each repo's CONTRIBUTING.md
  • Open issues/PRs — follow each repo's contribution guidelines

Community


License

Unless noted otherwise, HOL software is released under Apache-2.0. Check each repository's LICENSE for details.

Pinned Loading

  1. hcs-improvement-proposals hcs-improvement-proposals Public

    Hashgraph Online Website & Open Source Standards.

    TypeScript 26 16

  2. standards-sdk standards-sdk Public

    The official HOL Standards SDK, implementing the standards found in https://hol.org/docs/standards

    TypeScript 776 17

  3. desktop desktop Public

    HOL Desktop

    TypeScript 10 4

  4. awesome-codex-plugins awesome-codex-plugins Public

    A curated list of awesome OpenAI Codex plugins, skills, and resources. The #1 Codex Marketplace. See live plugins at: https://hol.org/registry/plugins

    Python 606 208

  5. hol-guard hol-guard Public

    AI antivirus for developer agents: protect Codex, Claude Code, Cursor, Gemini, OpenCode, plugins, skills, MCP servers, and AI harnesses before tools run.

    Python 388 8

Repositories

Showing 10 of 49 repositories

Top languages

Loading…

Most used topics

Loading…