Skillpack helps teams turn AI skills into trusted local agents that can run in their own environment and be used directly from Slack and Telegram. Our vision is to achieve distributed intelligence network, much like cremini mushrooms that grow from a vast, interconnected mycelial network.

skillpack.sh is an open-source way to package AI skills into runnable local agents. If skills and tools are like LEGO pieces, a SkillPack is the finished product that assembles them into a complete solution. Instead of juggling prompts, scripts, docs, and one-off automations, Skillpack gives you a simple way to:

• package AI skills into reusable agents
• run them locally
• keep sensitive data in your own environment
• use agents from tools your team already uses, like Slack and Telegram

Skillpack is built for teams that want AI Agents to be deployable, trusted, and easy to use.

1. Download the example

• Garry Tan SkillPack
• Company Deep Research SkillPack

2. Unzip it and Run ./start.sh on Mac OS, Or double click start.bat on Windows (see below), the server starts and opens http://127.0.0.1:26313 in your browser

# macOS / Linux
./start.sh

# Windows
start.bat

3. Enter an LLM API key (OpenAI or Claude API Key) in the left menu, use the prompt example to try it!
4. (Optional) Refer to the instructions Slack/Telegram Integrations below to integrate with Slack and Telegram.

npx @cremini/skillpack create

Step by step:

1. Set the pack name and description.
2. Add skills from GitHub repos, URLs, or local paths.
3. Add prompts to tell the agent how to orchestrate those skills.
4. Optionally package the result as a zip immediately.

# From a local file
npx @cremini/skillpack create --config ./skillpack.json

# From a remote URL (no directory = current directory)
npx @cremini/skillpack create comic-explainer --config https://raw.githubusercontent.com/CreminiAI/skillpack/refs/heads/main/examples/comic-explainer.json

Ready to run using "Run a skillpack" part

npx @cremini/skillpack zip

Produces <pack-name>.zip in the current directory.

When adding skills through create, the source accepts:

# GitHub shorthand
vercel-labs/agent-skills --skill frontend-design

# Full GitHub URL
https://github.com/JimLiu/baoyu-skills/tree/main/skills --skill baoyu-comic

# Local path
./skills/my-local-skill

Multiple skill names from the same source can be listed comma-separated.

The archive produced by zip is intentionally lightweight:

<pack-name>/
├── skillpack.json       # Pack configuration
├── AGENTS.md            # Optional pack policy
├── SOUL.md              # Optional pack persona
├── skills/              # Installed skills
├── start.sh             # One-click launcher for macOS / Linux
└── start.bat            # One-click launcher for Windows

The start scripts use npx @cremini/skillpack run . so Node.js is the only prerequisite — no pre-bundled server directory is included.

If present, AGENTS.md and SOUL.md are read by SkillPack itself when a new chat session starts. SkillPack injects them into the runtime system prompt as pack-level policy and persona, without depending on the host machine's AGENTS.md, .pi/SYSTEM.md, or APPEND_SYSTEM.md.

Talk to your Agents on Slack and Telegram

https://skillpack.gitbook.io/skillpack-docs/getting-started/slack-integration

https://skillpack.gitbook.io/skillpack-docs/getting-started/telegram-integration

The main use case is to run local agents on your computer and integrate them with Slack or Telegram so they can work for you and your team — operating entirely on your machine to keep all team data local and private, while continuously improving by learning new skills. Each SkillPack organizes skills around a well-defined job — for example: research a company by gathering information from multiple sources and produce a PowerPoint presentation from the findings.

Download Company Deep Research and try it! More examples can be found at skillpack.sh

Join our Discord at https://discord.gg/nj8Br4ePJc

MIT