Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/org-quicko/skillset/llms.txt

Use this file to discover all available pages before exploring further.

This guide walks you through deploying a Skillset Registry on your own machine, creating a Superadmin account, publishing your first Skill, and installing it into a coding agent. By the end you will have a working Registry and a Skill ready to use.

​
Prerequisites

Before you begin, make sure you have:
  • Docker (with the compose plugin) — to run the app and Postgres
  • Node.js 18+ and npm — to install and run the CLI
  • git — required by the CLI when installing or publishing from repository URLs
1

Clone the repository and configure your environment

Clone Skillset and copy the example environment file:
git clone https://github.com/org-quicko/skillset.git
cd skillset
cp .env.example .env
Open .env and fill in the required values. The defaults for DATABASE_URL and Postgres credentials already match what docker-compose.yml provisions. If you are using a local MinIO instance for object storage, the STORAGE_* defaults point to it; otherwise update them to your S3 bucket. The two variables you must set regardless are:
# Signing secret for sessions — generate a fresh one:
BETTER_AUTH_SECRET=   # openssl rand -base64 32

# Absolute URL this Registry is reached at (no trailing slash):
PUBLIC_URL=http://localhost:3000
Generate a secure BETTER_AUTH_SECRET in one command:
openssl rand -base64 32
2

Start the Registry

docker compose up
This starts two containers:
ContainerWhat it does
appRuns the API and serves the web UI at http://localhost:3000
postgresStores all Resource metadata; migrations run automatically on startup
Wait for the log line Listening on http://0.0.0.0:3000 before continuing. The app waits for Postgres to be ready and runs all migrations before it starts serving traffic.
The first startup creates the pg_trgm Postgres extension, which powers fuzzy search. On most managed Postgres instances this requires a privileged role. If the migration fails, have a DBA run CREATE EXTENSION IF NOT EXISTS pg_trgm; against your database once, then restart.
3

Create your Superadmin account

Open http://localhost:3000 in your browser. You will be prompted to create the Superadmin account — the first User to sign up is the permanent Superadmin, with full control over the Registry.Fill in your email address, name, and a password, then submit the form. You will be logged in immediately.
There is exactly one Superadmin per Registry. The role is permanent and cannot be transferred. Use an email address you control long-term.
4

Install the CLI

In a new terminal, install the skillset CLI globally:
npm install -g @in-org-quicko/skillset-cli
Verify the install:
skillset --version
5

Log in to your Registry

skillset login --registry http://localhost:3000
The CLI will prompt for your email and password (the Superadmin credentials you just created). On success it stores a Token for that Registry in your local config.
skillset whoami
# → Logged in as you@example.com (superadmin) on http://localhost:3000
6

Create and publish your first Skill

Create a new directory for your Skill. Every Skill must have a SKILL.md at its root — the name and description in its frontmatter are the source of truth for the Resource.
mkdir code-review
cd code-review
Create SKILL.md:
---
name: code-review
description: Reviews code for common issues and best practices
---

## Overview

This skill reviews a code diff or file for common issues including:

- Logic errors and off-by-one mistakes
- Missing error handling
- Security concerns (injections, exposed secrets, over-broad permissions)
- Style and readability problems

## Usage

Pass the code to review as context. Ask the agent to review it for issues,
suggest improvements, or check it against a specific standard.
Publish the Skill from the current directory:
skillset publish .
# → Published code-review (1 file) to http://localhost:3000
7

Install the Skill into a coding agent

From any project directory, install the Skill you just published:
skillset install code-review
The CLI downloads the Skill’s Artifact, writes it to .agents/skills/code-review/, creates or updates skillset-lock.json at the project root, and symlinks the chosen agent’s own skills directory to .agents/skills/.Check that the install is current:
skillset list
# → code-review   current
skillset list reads the Lockfile (skillset-lock.json) to compare what’s on disk against the Registry’s latest version. A Skill you’ve edited locally shows modified; one the Registry has updated since you installed it shows outdated.

​
What just happened

You deployed a full Skillset Registry (app + Postgres), created a Superadmin account, published a Skill with the CLI, and installed it into your project. Your Registry is now ready for your team.

CLI Commands

Every skillset command with all its flags — search, info, update, remove, and more.

MCP Server

Let your coding agent search and install Skills directly, without the CLI.

Deployment

Configure Postgres schemas, swap MinIO for S3, and harden for production.

Concepts

Namespaces, Lockfiles, Imports, Submissions, and the rest of the data model.

Build docs developers (and LLMs) love

Get started for freeTalk to us