agentregistry CLI
A registry and arctl CLI for MCP servers, agents, skills and prompts: publish, run and configure clients in one place.
- 540 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- Installation runs a script downloaded through `curl | bash`.
- The registry runs containers and deploys servers in your infrastructure.
Install
- Works in
Manual install
curl -fsSL https://raw.githubusercontent.com/agentregistry-dev/agentregistry/main/scripts/get-arctl | bashThe first README quick start command. Docker Desktop with Docker Compose v2 is required.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Installation runs a script downloaded through `curl | bash`.
- The registry runs containers and deploys servers in your infrastructure.
This is third-party code. Review the repository files before installing.
What it does
agentregistry keeps MCP servers, agents, skills and prompts in one catalog. Servers are registered from npm, PyPI, OCI images or remote HTTP/SSE endpoints, and each entry has versions and environment variable descriptions. The `arctl` command scaffolds artifacts (`arctl init skill`, `arctl init agent`), builds and publishes images (`arctl build ./skill --push`) and runs them locally or in Kubernetes through Helm. `arctl configure cursor` configures an IDE, and the web UI on `localhost:12121` shows the catalog.
Who it is for. Platform teams and developers maintaining a shared catalog of vetted MCP servers and skills for a company.
Good fit when
- MCP servers are scattered across npm, PyPI and Docker Hub, and a single list of approved ones is needed.
- A team publishes its own skill and hands it to colleagues with one command.
- An agent with its servers and skills is packaged into one deployable unit.
- A developer connects Cursor or Claude Desktop to approved servers through `arctl configure`.
- The same registry runs locally on a laptop and in Kubernetes.
Not a fit when
- You need a public read-only catalog, and for a personal set of servers a registry is overkill.
- There is no Docker Desktop with Compose v2 and no Kubernetes, while the quick start relies on them.
- A list of links in a README without versions or image builds is enough.
Example request
Scaffold a skill with `arctl init skill`, build the image and publish it to the registry.Limitations
The quick start requires Docker Desktop with Docker Compose v2 and installs the CLI with a `curl | bash` script. The registry brings up its own PostgreSQL database. It is infrastructure for teams and rarely needed by a solo user. The last commit dates to October 2026.
How to disable. Run `docker compose -f agentregistry-compose.yml down` and the data stays. For full removal add `-v` and delete the `arctl` binary.
FAQ
What gets published to the registry?
MCP servers, agents, skills and prompts. Servers can come from npm, PyPI, OCI images and remote HTTP/SSE endpoints.
Which clients does `arctl` configure?
The README names Claude Desktop, Cursor, Kiro IDE and VS Code.
Is Kubernetes required?
No, the registry runs locally through Docker Compose. Kubernetes is for shared deployment through Helm.
How do I stop the registry without losing data?
With `docker compose -f agentregistry-compose.yml down`.
What is agentgateway in this setup?
A reverse proxy for MCP traffic: clients connect to one address while the gateway routes calls and checks access.
Discussion
No questions or reviews yet.
Related
CLI toolsOpen registry format for skills and agents plus the aireg CLI: pinned versions, rollback and sync for Claude Code, Codex, Cursor.
A CLI that installs shared skills, commands, subagents, hooks and MCP servers into eight AI coding tools in their native layouts
MCP Gateway Registry
mcp-gateway-registry
One registry and gateway for MCP servers, agents and skills: shared login, OAuth, audit, OpenTelemetry. Community project for Docker and cloud
A self-hosted MCP and A2A gateway with identity provider integration, access control and observability for enterprise agents
MCP server for Kubernetes: the agent reads cluster state, diagnoses pod failures and runs operations through kubectl
AWS's official set of MCP servers, skills and plugins so an agent can build, deploy and operate on AWS
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| agentregistry (this card) | no | 540 | ||
| MCP Registry | no | 7,464 | a catalog of MCP servers; agentregistry adds agents, skills, prompts, image builds and Kubernetes deployment | |
| MCP Gateway Registry | no | OAuth | 1,048 | a gateway and registry for MCP; agentregistry focuses on publishing and delivering artifacts through the `arctl` CLI |
README in short
The README describes a registry for MCP servers, agents, skills and prompts, the `arctl` CLI, a web UI, Kubernetes deployment and the agentgateway pairing.