registry-broker-skills
Health Gecti
- License — License: Apache-2.0
- Description — Repository has a description
- Active repo — Last push 0 days ago
- Community trust — 197 GitHub stars
Code Uyari
- process.env — Environment variable access in examples/ledger-auth.js
- network request — Outbound network request in examples/ledger-auth.js
- process.env — Environment variable access in examples/register-agent.js
- network request — Outbound network request in examples/register-agent.js
- process.env — Environment variable access in examples/search-and-chat.js
- network request — Outbound network request in examples/search-and-chat.js
- process.env — Environment variable access in index.js
- network request — Outbound network request in index.js
Permissions Gecti
- Permissions — No dangerous permissions requested
This MCP server provides AI agent skills for a Universal Agentic Registry. It allows AI assistants to search live agent inventories, chat with registered agents, and register their own agents across multiple protocols.
Security Assessment
The tool makes outbound network requests to the public registry API in both its main `index.js` file and its example scripts. It relies on environment variables to handle credentials (such as ledger authentication), which is a standard and relatively safe practice, provided you do not hardcode sensitive data in your `.env` files. There are no hardcoded secrets, no dangerous permissions requested, and no shell command executions detected. Overall risk: Low.
Quality Assessment
The project demonstrates solid maintenance and community traction. It received updates as recently as today, and has garnered nearly 200 GitHub stars. It uses a standard permissive license (Apache-2.0) and features professional documentation, including a live OpenAPI spec and DOI reference.
Verdict
Safe to use — it is an actively maintained, well-documented tool that interacts safely with external APIs.
AI agent skills for the Universal Registry - search, chat, and register 72,000+ agents across 14+ protocols. Works with Claude, Codex, Cursor, OpenClaw, and any AI assistant.
HOL Registry
![]() |
AI agent skills for the Universal Agentic Registry. Search live AI agent inventory, chat with any agent, register your own — consumable by Claude, Codex, Cursor, OpenClaw, and any AI coding assistant. Live Registry | API Docs | SDK Docs |
|---|
What is this?
This repository contains skill definitions for the Universal Agentic Registry — the connectivity layer for the autonomous web. Skills are instruction files that teach AI coding assistants how to interact with the Registry Broker API.
It is designed for developer-intent discovery around:
- MCP server registry and MCP server discovery
- Agent skills registry and skill package publishing
- Agent directory CLI workflows and API-first integration
The SKILL.md file can be consumed by:
- Claude Code / Claude Desktop — via MCP or direct skill loading
- OpenAI Codex / ChatGPT — as context instructions
- Cursor — as project instructions
- OpenClaw / ClawHub — native skill format
- Any AI coding assistant — universal markdown format
Share & Embed
Use these snippets in your README, docs, or release notes to link canonical skill surfaces:
[](https://hol.org/registry/skills)
[](https://github.com/hashgraph-online/skill-publish)
Direct references:
- Skills index:
https://hol.org/registry/skills - Product docs:
https://hol.org/docs/registry-broker/ - Interactive API docs:
https://hol.org/registry/docs - OpenAPI schema:
https://hol.org/registry/api/v1/openapi.json - Live stats endpoint:
https://hol.org/registry/api/v1/dashboard/stats - Skill publishing action:
https://github.com/hashgraph-online/skill-publish
Schema & Validation
- Canonical manifest schema:
schemas/skill.schema.json - SchemaStore submission kit:
references/SCHEMASTORE-SUBMISSION.md - Syndication kit:
references/SYNDICATION-KIT.md - APIs metadata:
apis.json - LLM index:
llms.txt
DOI Readiness
- Zenodo metadata:
.zenodo.json - Citation metadata:
CITATION.cff
What is the Universal Registry?
One standards-compliant API to access live AI agent inventory from:
| Protocol | Description |
|---|---|
| AgentVerse | Fetch.ai autonomous agents |
| Virtuals | Tokenized AI agents |
| A2A | Google's Agent-to-Agent protocol |
| MCP | Anthropic's Model Context Protocol |
| ERC-8004 | On-chain agent verification |
| x402 Bazaar | Agent payment rails |
| OpenRouter | LLM gateway |
| NANDA | Decentralized AI |
| Near AI | Near Protocol agents |
| OpenConvAI | Conversational AI standard |
| XMTP | Decentralized messaging |
| ANS | Agent Name Service |
| PulseMCP | MCP server registry |
| HCS-10 | Hedera Consensus Service agents |
Quick Start
Option 1: NPX CLI (quickest)
# Search for agents
npx @hol-org/registry search "trading bot"
# Get platform statistics
npx @hol-org/registry stats
# Resolve a UAID to agent details
npx @hol-org/registry resolve uaid:aid:fetchai:agent123
# Start a chat (requires API key, uses XMTP by default)
export REGISTRY_BROKER_API_KEY="your-key"
npx @hol-org/registry chat uaid:aid:fetchai:agent123 "Hello!"
# Verify Moltbook agent ownership (required to send as agent; also unlocks broker registration)
MOLTBOOK_API_KEY="mb_xxxxx" npx @hol-org/registry claim
# Mark a verified Moltbook agent as "registered" in the broker (directory benefits)
npx @hol-org/registry register uaid:aid:moltbook:yourAgent --description "Updated description"
# Check credit balance
npx @hol-org/registry balance
# Skill registry (publish/find decentralized skills)
export REGISTRY_BROKER_API_KEY="your-key"
npx @hol-org/registry skills config
npx @hol-org/registry skills init --dir ./my-skill --name "my-skill" --version 0.1.0
npx @hol-org/registry skills lint --dir ./my-skill
npx @hol-org/registry skills list --name "my-skill" --limit 5
npx @hol-org/registry skills verify --name "my-skill" --tier basic --account-id 0.0.1234
npx @hol-org/registry skills verification-status --name "my-skill" --account-id 0.0.1234
npx @hol-org/registry skills quote --dir ./path/to/skill --account-id 0.0.1234
npx @hol-org/registry skills publish --dir ./path/to/skill --account-id 0.0.1234
npx @hol-org/registry skills job <jobId> --account-id 0.0.1234
# Get skill file URL
npx @hol-org/registry skill
Option 2: Use the Skill File
Copy SKILL.md to your project or reference it in your AI assistant's context.
Option 3: MCP Server (recommended for Claude/Cursor)
npx @hol-org/hashnet-mcp up --transport sse --port 3333
Option 4: TypeScript SDK
npm install @hashgraphonline/standards-sdk
import { RegistryBrokerClient } from "@hashgraphonline/standards-sdk";
const client = new RegistryBrokerClient();
// Search for AI agents
const results = await client.search({ q: "trading bot" });
// Chat with an agent
const session = await client.createChatSession({ uaid: "uaid:aid:..." });
const response = await client.sendMessage({
sessionId: session.sessionId,
message: "Hello!"
});
Option 5: Direct API (curl)
# Get API key at https://hol.org/registry
export REGISTRY_BROKER_API_KEY="your-key"
# Search for agents
curl "https://hol.org/registry/api/v1/search?q=trading+bot&limit=5"
# Create chat session
curl -X POST "https://hol.org/registry/api/v1/chat/session" \
-H "Content-Type: application/json" \
-H "x-api-key: $REGISTRY_BROKER_API_KEY" \
-d '{"uaid": "uaid:aid:..."}'
Repository Contents
registry-broker-skills/
├── SKILL.md # Universal skill definition (main instruction file)
├── README.md # This file
├── LICENSE # Apache-2.0
├── scripts/ # Helper bash scripts
│ ├── search.sh # Search for agents
│ ├── chat.sh # Start a chat session
│ ├── resolve.sh # Resolve UAID to details
│ ├── stats.sh # Get platform statistics
│ └── balance.sh # Check credit balance
├── examples/ # Code examples
│ ├── search-and-chat.js # Search and chat workflow
│ ├── register-agent.js # Agent registration
│ └── ledger-auth.js # Wallet authentication
├── references/ # API documentation + submission kits
│ ├── API.md # Complete API reference
│ ├── PROTOCOLS.md # Supported protocols
│ ├── MCP.md # MCP server reference
│ ├── SCHEMASTORE-SUBMISSION.md
│ └── SYNDICATION-KIT.md
└── schemas/
└── skill.schema.json # JSON Schema for skill.json
Scripts
Quick bash scripts for common operations:
# Search for agents
./scripts/search.sh "trading bot" 5
# Resolve a UAID
./scripts/resolve.sh "uaid:aid:fetchai:agent123"
# Start a chat session (requires API key)
export REGISTRY_BROKER_API_KEY="your-key"
./scripts/chat.sh "uaid:aid:fetchai:agent123" "Hello!"
# Check platform stats
./scripts/stats.sh
# Check credit balance
./scripts/balance.sh
API Capabilities
Discovery
- Keyword Search — Find agents by name, description, capabilities
- Vector/Semantic Search — Natural language agent discovery
- Capability Search — Filter by specific agent capabilities
- Agent Details — Full profile, metadata, trust scores
- Similar Agents — Find related agents
Chat
- Create Session — Start conversation with any agent
- Send Messages — Real-time chat with streaming support
- History — Retrieve conversation history
- Compact — Summarize history for context window management
Registration
- Quote — Get credit cost estimate
- Register — Add your agent to the registry
- Update — Modify agent profile
- Unregister — Remove agent
Credits & Payments
- Balance — Check credit balance
- HBAR Payments — Purchase credits with HBAR
- Stripe Payments — Purchase credits with card
- X402 (EVM) — Purchase credits via EVM wallets
Authentication
- API Key — Traditional API key auth
- Ledger Auth — Wallet-based authentication (Hedera, Ethereum, Base, Polygon)
API & Documentation
| Resource | Link |
|---|---|
| Live Registry | hol.org/registry |
| API Documentation | hol.org/docs/registry-broker |
| Interactive API Docs | hol.org/registry/docs |
| SDK Documentation | hol.org/docs/libraries/standards-sdk |
| Postman Collection | Run in Postman |
| Insomnia | Import OpenAPI |
| OpenAPI Spec | openapi.json |
| Live Stats Endpoint | dashboard/stats |
Related Repositories
| Repository | Description |
|---|---|
hashnet-mcp-js |
MCP server for Registry Broker |
standards-sdk |
TypeScript/JavaScript SDK |
universal-registry-quickstart |
Quickstart example project |
skill-publish |
GitHub Action for quote/publish/poll skill release workflows |
Using with AI Assistants
Claude Code / Claude Desktop
Add to your Claude configuration:
{
"mcpServers": {
"hashnet": {
"command": "npx",
"args": ["@hol-org/hashnet-mcp@latest", "up", "--transport", "stdio"]
}
}
}
Or reference SKILL.md directly in your conversation.
Cursor
Copy SKILL.md to your project root, or add to .cursor/rules/.
OpenClaw / ClawHub
The SKILL.md file uses OpenClaw's native YAML frontmatter format and is directly compatible.
Codex / ChatGPT
Reference the skill file in your system prompt or paste its contents as context.
Score HOL Points
Contribute to this repository and score HOL Points!
- Fix bugs or improve documentation
- Add new features or examples
- Submit pull requests to score points
Points can be used across the HOL ecosystem. Learn more
License
Apache-2.0
Cite this repository
If you reference this project in documentation, reports, or research, use the metadata in CITATION.cff.
CI/CD: Publish from GitHub
This repository supports one-step HCS-26 publishing from GitHub using the first-party action hashgraph-online/skill-publish@v1.
- Workflow:
.github/workflows/skills-publish.yml - Triggered on
release: published,pushtomain, and manual dispatch - Automatically validates package files, gets quote, publishes, polls job status, and posts on-chain references back to release notes/PR comments
- Automatically stamps
repoandcommitmetadata into the published skill payload
Required repository secret:
RB_API_KEY
Yorumlar (0)
Yorum birakmak icin giris yap.
Yorum birakSonuc bulunamadi
