packs

Documentation

Learn how to install and use packs to supercharge your AI agents.

One pack. Your agent goes from "I'll attempt to debug this" to following a proven, systematic process. Better results, less fumbling.

Installation

Install the packs CLI using Homebrew.

Supports macOS and Linux. The CLI is a single binary with no dependencies.

Usage

The CLI provides a few simple commands to discover and install packs.

Browse

Interactive

Launch the TUI to explore packs visually.

Search

Find packs by keyword, tag, or description.

Install

Fetch a pack and output its contents. Pipe to your agent or save to a file.

Inspect

View metadata and details about a pack.

GitHub Fetch

Fetch packs directly from any GitHub repository—public or private.

Tip: Run packs login to access private repositories.

Pack Types

Packs come in two types, each serving a different purpose.

Skills

SKILL.md

Step-by-step workflows and processes. How to debug, how to write commits, how to review code.

Contexts

CONTEXT.md

Reference material and best practices. Framework patterns, API conventions, language idioms.

Creating Packs

A pack is just a folder with two files. Simple by design.

my-pack/
├── pack.yaml
└── SKILL.md

pack.yaml

name: my-pack
version: 1.0.0
type: skill
description: One line describing what this does
author: your-username
tags:
  - relevant
  - tags
license: MIT

Content File

Your SKILL.md or CONTEXT.md should include:

  • When to use this pack
  • Clear instructions or reference material
  • Working examples
  • Common pitfalls to avoid

Submit a Pack

Share your pack with the community in one command.

Or submit a pull request directly to the packs-registry.

Authentication

Authenticate with GitHub to access private repos and submit packs.

Sign in with GitHub

Sign out

Check current user