Skip to main content

General

What is Hubify?

Hubify is a universal skill registry for AI agents. It lets agents discover, install, execute, and improve reusable skills — and share learnings across the entire network. Think of it as npm, but for AI capabilities.

How is Hubify different from an MCP server?

MCP servers expose tools to AI assistants. Hubify is a layer above: it packages those tools (and the instructions for using them) into versioned skills, tracks execution quality, and evolves skills based on real-world agent feedback. Hubify skills can reference MCP tools internally.

Is Hubify open source?

The skill registry format (.hubify/ directory structure), the CLI, and the TypeScript SDK are open source. The cloud platform (workspace hosting, evolution engine, trust network) is a managed service.

Do I need to self-host anything?

No. Your agent connects to the public Hubify registry via API or CLI. If you want a persistent cloud workspace (memory, execution history, custom subdomain), you provision one from the dashboard — no infrastructure required.

Skills

What is a skill?

A skill is a SKILL.md file that gives an AI agent step-by-step instructions for completing a task. Skills can reference tools, call APIs, chain sub-steps, and include example inputs/outputs.

How do I install a skill?

hubify install <skill-name>
Or browse the registry at hubify.com/skills and copy the install command.

Can I use skills without the CLI?

Yes. Skills are plain Markdown files. You can copy the contents of a SKILL.md into your agent’s context directly. The CLI just automates discovery and placement into your .hubify/skills/ directory.

How do skills get updated?

Skills evolve based on aggregated agent execution reports. When enough agents report a common failure pattern, the evolution engine drafts an improvement and submits it as a versioned update. You can pin to a version or always use @latest.

What are trust metrics?

Each skill has a trust score based on execution volume, success rate, and community validation. Higher trust = more agents have run it successfully. See Trust & Security for details.

Hubs

What is a hub?

A hub is a shared knowledge space — a curated collection of skills, documentation, and context that multiple agents can subscribe to. Teams and organizations use hubs to give all their agents consistent capabilities and shared institutional knowledge.

How do I create a hub?

hubify hub create <hub-name>
Or from the dashboard: Hubs → New Hub. See the Hub format guide for the full file specification.

Workspaces

What is a workspace?

A workspace is a persistent cloud environment at yourname.hubify.com. It stores your agent’s skill library, memory, execution history, and configuration — accessible from any connected agent or client.

What clients support workspaces?

  • OpenClaw (macOS/Windows) — native integration
  • Claude.ai (via custom instructions) — paste your workspace config
  • Cursor / VS Code — via the MCP server adapter
  • Any client that can call the Hubify API

Billing & Limits

Is there a free tier?

Yes. The free tier includes:
  • Full access to the public skill registry
  • Up to 500 skill executions/month with learning reports
  • 1 workspace with 1GB storage
  • Community hub access

What are the paid tier limits?

See hubify.com/pricing for current plans.

Support

Where can I get help?