Agentic Resource Registry

Agentic Resource Registry

Keeps a team's coding agent skills, hooks, prompts and agent files in one self-hosted registry, with review of new entries and usage insight. Suits engineering teams that run several coding agents.

Made by AgenticResourceRegistry on GitHub. These notes are Mindket's, written from its README; the project's own files and license are in its repository.

About

What it is: Agentic Resource Registry is a self-hosted registry where an organization keeps the skills, agent files, MCP clients, hooks, prompts and sandboxes its coding agents use. It was called Observal before. Teams get one place to find what already exists, plus usage insight that shows which resources fail quietly.

Who it suits: Engineering teams that share coding agent resources across many developers get the most from it. So do the people who write those resources and the ones who approve them.

Industry | Who uses it | How they use it | Example
Software companies | Developer tooling teams | Keep internal skills, agent files and MCP clients in one registry | A team publishes a code review skill so other developers install it instead of writing a second one
Large engineering organizations | Reviewers and approvers | Check new submissions, approve internal agents and compare version diffs | A reviewer reads what changed between two versions of an agent before approving it
Teams on mixed tooling | Developers | Install a resource once and get the right config for the coding tool each person uses | One developer on Cursor and another on Claude Code use the same skill

A solo developer who keeps a few skills in one folder can skip it.

Where to use it:

• Claude Code, Cursor and Codex
• Kiro, Pi and Copilot
• OpenCode and other coding tools
• Your own server, since the registry is self-hosted

Use case | Who | What they get back
Find existing resources | Developers | A shared place to look before writing a new skill or agent file
Review submissions | Maintainers | A queue to approve internal agents and inspect version diffs
Set up another tool | Developers | The correct config for each supported coding tool, not separate setup notes
Learn from usage | Authors | Insight into usage patterns that turns silent failures into feedback

For newcomers and experts: It is one shared place for the skill files, agent files and tool settings your coding agent reads. One person publishes a resource and others install it, so nobody writes a second copy. A first small step is to list the resources your team keeps in separate repositories today.

Underneath, it is a registry with a command line client, observal-cli. It adds a review step and version diffs for submissions, and it renders one resource into the config each supported coding tool expects. Coding agents often give subtly wrong answers without raising an error, so usage insight is how authors learn what went wrong. The trade-off is that you host and maintain the server yourself.

Authors publish a resource
Reviewers approve it
Developers install it from one place
Usage insight returns to the authors

Good to know:

• Self-hosted: you run the registry yourself.
• License: Apache 2.0.
• The command line client is observal-cli on PyPI and needs Python 3.11 or newer.
• The command line tool is still named observal.

On GitHub, 7 November 2026

• Stars: 4,320
• Forks: 1,142
• Watching: 12
• Open issues and pull requests: 271
• Language: Python
• License: Apache-2.0
• Topics: agents, analytics, antigravity, claude-code, cli-tool, codex, cursor, cursor-ai, insights, kiro, large-language-models, mcp

Links

• Code: https://github.com/AgenticResourceRegistry/AgenticResourceRegistry
• Website: https://agenticresourceregistry.io/
• Releases: https://github.com/AgenticResourceRegistry/AgenticResourceRegistry/releases
• Issues: https://github.com/AgenticResourceRegistry/AgenticResourceRegistry/issues
• Star history: https://www.star-history.com/#AgenticResourceRegistry/AgenticResourceRegistry&Date

Get the code

git clone https://github.com/AgenticResourceRegistry/AgenticResourceRegistry.git

Then follow the setup in its README. Nothing in this kit runs by itself.