The universal registry for AI agent capabilities

OpenBooklet — one registry
for every AI agent

OpenBooklet is the open registry for AI agent skills, workflows, MCP servers, and collections. Publish once — any agent discovers and uses them instantly. Free, open, verified.

# Single-task instructions for any AI agent
skills:
- openbooklet.com/s/code-review
- openbooklet.com/s/security-audit@^2.0
- openbooklet.com/s/accessibility-check@latest
# Auto-converted to your agent's native format
Free
Forever
<50ms
Global latency
All
Agent formats
Verified
Tamper-proof ownership

const capabilities = "Everything your agent needs"

Four primitives. One registry. Any agent.

Skills

Single-task instructions — the atomic building blocks of agent behavior. Versioned, verified, and auto-converted to your agent's native format.

skills:
  - openbooklet.com/s/code-review
  - openbooklet.com/s/security-audit@^2.0
Browse Skills

Workflows

Multi-step skill chains with SHA-256 locked dependencies. One request fetches everything — deterministic, every time.

lint → test → review → deploy
Browse Workflows

MCP Servers

Pre-configured server definitions agents discover and connect to instantly. 17 tools, 7 resources, zero setup.

npx -y @openbooklet/mcp-server
Browse MCP Servers

Collections

Curated sets of skills, workflows, and MCP servers — grouped by topic, team, or use case. Public or private. Subscribe once, stay current.

openbooklet.com/c/frontend-essentials
Browse Collections

publish().then(discover)
.then(consume)

From publishing to agent consumption in three steps.

01

Publish

Create a skill, workflow, or MCP server. We handle validation, versioning, and format conversion automatically.

02

Discover

Search by keyword or describe what you need in natural language. AI-powered search finds the best match across the entire registry.

03

Consume

Fetch via URL, API, MCP, CLI, or SDK. Auto-converted to your agent's native format. Cached globally, delivered in under 50ms.

type Registry = Verified & Open & Free

Every capability designed for AI-first consumption with human oversight.

12 Agent Formats

Write once, auto-convert to Claude Code, Cursor, Windsurf, Copilot, GPTs, LangChain, and more.

AI-Powered Discovery

Describe what you need in natural language. Semantic search finds the best match — not just keywords.

Workflow Engine

Chain skills into multi-step pipelines. Dependencies are locked and delivered as a single bundle.

MCP Servers

Pre-configured server definitions agents can discover and connect to. Extend capabilities across the registry.

Verified Ownership

Every asset is content-hashed and timestamped at publish. Permanent, tamper-proof authorship proof.

Safety First

Automated safety scanning catches prompt injection, data exfiltration, and malicious patterns before publishing.

Global Edge Delivery

330+ edge locations worldwide. Sub-50ms response times. Smart caching with ETag validation.

Semantic Versioning

Pin versions, auto-update within ranges, view changelogs, and roll back from the dashboard. Dependencies are SHA-256 locked.

const { workflows, collections } = registry

Workflows for multi-step automation. Collections for curated grouping.

Workflows

Multi-step skill chains with locked dependencies. One request fetches everything — deterministic, every time.

openbooklet.com/workflows
deploy-pipeline
DevOps6 steps
full-code-review
Quality4 steps
llm-eval-suite
AI/ML9 steps
Browse Workflows

Collections

Personal or team-curated sets. Mix skills, workflows, and MCP servers. Public or private — your choice.

openbooklet.com/collections
my-agent-starter-kit
6 items
team-code-review-set
4 items
deploy-pipeline-bundle
9 items
Browse Collections

await registry.publish(skill)

Publish once, reach every AI agent on Earth. Your reputation grows with every pull.

Publisher Dashboard
AK
alex-kim✦ MASTER
OB-K4X92
12,847
Total pulls
23
Skills
#1
Leaderboard
Pulls — last 30 days +34%
code-review-pro
4.95,421 pulls
security-audit
4.83,102 pulls

Verified trust tiers

Progress from New → Rising Star → Pro → Master as your contributions earn recognition across the ecosystem.

Real-time analytics

Pull counts, agent compatibility scores, and rating distributions. Always know exactly who's using your work.

Global leaderboard

Top publishers surface organically. Get discovered by agents and developers building the next generation of AI.

import { URL, API, MCP, CLI, SDK } from "@openbooklet"

Six ways to integrate. From zero-config URL fetching to deep MCP tooling. All open, all documented.

Direct URL

Any agent that can fetch a URL — zero setup.

Docs
# In your CLAUDE.md or agent config
skills:
  - openbooklet.com/s/code-review
  - openbooklet.com/s/security-audit@^2.0

REST API

70+ endpoints for search, fetch, publish, and analytics.

Docs
curl https://openbooklet.com/api/v1/skills/code-review-pro \
  -H "Authorization: Bearer ob_live_..."

MCP Server

17 tools + 7 resources for native AI agent integration.

Docs
{
  "mcpServers": {
    "openbooklet": {
      "command": "npx",
      "args": ["-y", "@openbooklet/mcp-server"]
    }
  }
}

CLI

Search, pull, install, publish from your terminal.

Docs
$ npm install -g @openbooklet/cli
$ ob search "react patterns"
$ ob pull react-patterns-2026
$ ob publish my-skill.md

TypeScript SDK

Zero-dependency client library.

Docs
import { OpenBooklet } from "@openbooklet/sdk";

const ob = new OpenBooklet({ apiKey: "ob_live_..." });
const results = await ob.search("security audit");

GitHub Import

Import skills from any public repo or raw URL.

Docs
# Import from a raw URL
$ ob import https://raw.githubusercontent.com/  you/repo/main/SKILL.md

# Or scan a repo for SKILL.md files
$ ob import github.com/you/your-repo

The universal registry for AI agent capabilities

Publish skills, workflows, MCP servers, and collections. Reach every AI agent on Earth.

OpenBooklet — The Universal Registry for AI Agent Capabilities