SkillPack — Pack & Deploy Local AI Agents
Package AI skills into runnable local agents and deploy to your team in minutes, with Slack and Telegram integration.
- Source repo
- CreminiAI/skillpack
- Stars
- ★ 1.2k
- Last updated
- 1mo ago
- License
- MIT
- Primary language
- TypeScript
- FA score
- 40/100 · Major gaps
At a glance
- How it runs
- Works with
- Universal · cross-platformOpenAI API · Claude API
- Cost
- Free software; you pay for model usage
- Setup effort
- Low · running in minutes
- You'll need
- Typical use
- A team wants to package a company research skill into a local agent and deploy it on an internal server, accessible via Slack for all members.
- Not a fit if
- Teams unwilling to obtain Slack App/Bot or Telegram Bot tokens
- Users whose environment cannot install Node.js 22.19.0+
- Source review
- 40/100 · Major gaps 2 safety controls not found
What does this agent do, and when should you use it?
SkillPack is an open-source tool that packages AI skills into runnable local agents, enabling teams to deploy and run AI in their own environment while interacting via Slack or Telegram. It simplifies managing prompts, scripts, docs, and automations by turning them into reusable agents. SkillPack provides a CLI to create, package, and run agents, supporting skills from GitHub repos, URLs, or local paths. The generated zip is lightweight, requiring only Node.js 22.19.0+, and includes configuration, skills, and launch scripts. It also supports scheduled jobs via job.json and injects AGENTS.md and SOUL.md as pack-level policy and persona. SkillPack aims to build a distributed intelligence network, turning AI skills into trusted local agents.
SkillPack offers CLI commands (npx @cremini/skillpack create/zip/run) to create, package, and run agents. During creation, users set a pack name and description, add skills from GitHub repositories, URLs, or local paths, and provide prompts to orchestrate those skills. Running npx @cremini/skillpack run . starts a local server (default at http://127.0.0.1:26313) where users input an OpenAI or Claude API key. The output zip contains skillpack.json (configuration), skills/ folder, and start scripts (start.sh/start.bat). It integrates with Slack and Telegram, allowing team members to interact with agents via chat. It also supports scheduled tasks via job.json and reads AGENTS.md and SOUL.md to inject pack-level policy and persona at session start.
- A team wants to package a company research skill into a local agent and deploy it on an internal server, accessible via Slack for all members.
- A developer wants to quickly combine skills from GitHub repos (e.g., vercel-labs/agent-skills) to create a custom agent and share it with colleagues.
- An enterprise needs to keep sensitive data on-premises while allowing employees to use AI assistants through Telegram for everyday tasks.
- A team wants to encapsulate a recurring task (e.g., weekly report generation) into an agent that runs automatically on a schedule.
- A manager wants to consolidate prompts, scripts, and documents into a reusable agent and distribute it to the team without individual setup.
How do you install or deploy this agent?
SkillPack requires Node.js 22.19.0+. The CLI is used via npx, no global installation needed. For example packs, download the zip, unzip, and run ./start.sh (macOS/Linux) or double-click start.bat (Windows).
How do you use this agent?
- Create a new skill pack: run
npx @cremini/skillpack create, follow prompts to set pack name/description, add skills (GitHub URLs, local paths, etc.), and define prompts. - Run the agent: in the pack directory, run
./start.shorstart.bat, open http://127.0.0.1:26313, enter an OpenAI or Claude API key, and start using it. - Integrate with Slack or Telegram: follow the docs to get tokens and configure, then use agents in chat.
- Package for distribution: use
npx @cremini/skillpack zipto generate a zip that can be shared.
What are this agent's strengths and limitations?
- Turns AI skills into reusable agents, reducing deployment and maintenance effort.
- Supports importing skills from GitHub, URLs, and local paths for flexible composition.
- Offers Slack and Telegram integration for team-wide access in existing tools.
- Produces lightweight zip packages requiring only Node.js, no extra dependencies.
- Supports scheduled jobs and pack-level policy injection, suitable for enterprise use.
- Requires Node.js 22.19.0+ which may not be available in all environments.
- Needs an OpenAI or Claude API key from the user, incurring API costs.
- Slack integration requires obtaining App and Bot tokens, adding configuration overhead.
- At present, examples are limited and documentation may be incomplete.
How does this agent compare with similar options?
Compared to directly using OpenAI or Claude APIs, SkillPack focuses on packaging and deploying skills as agents with Slack/Telegram integration, offering a more team-oriented, local deployment approach rather than a single API call.
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| SkillPack — Pack & Deploy Local AI Agents This agent | 40 · Major gaps | CLIFree + model costs | ★ 1.2k | 1mo ago | TypeScript | OpenAI API · Claude API |
| Agent Swarm | 47 · Major gaps | Agent plugin / skillFree + model costs | ★ 904 | today | TypeScript | Codex · Claude Code |
| Echo Agent | 60 · Some gaps | CLIFree + model costs | ★ 1.1k | today | Python | — |
| Clawith — OpenClaw for Teams | 45 · Major gaps | Self-hosted serviceFree + model costs | ★ 4.2k | 1mo ago | Python | OpenAI API · Claude API |
How does FollowAgents rate this agent?
Why each dimension lost points
Evidence shows: README emphasizes local running and keeping sensitive data in user environment, but lacks specific implementation details for least privilege; user confirmation mechanism not mentioned; data flow transparency partially described (e.g., Slack/Telegram integrations); sensitive data handling mentioned but lacks concrete measures; dependency security not audited; external effects (e.g., Slack/Telegram integrations) described but side effects not addressed; rollback mechanism not mentioned; source attribution present via GitHub repo and author info. Deductions: lack of concrete implementation and documentation support.
Evidence shows: code and tests exist, but test coverage is limited; dependencies pinned but availability not verified; failure messages not documented. Deductions: incomplete tests, dependency availability unverified.
Evidence shows: README describes multiple use cases (local running, Slack/Telegram integrations), but capability boundaries not explicit; trigger precision (e.g., scheduling) mentioned but not detailed; environment fit (macOS/Windows/Linux) documented. Deductions: insufficient documentation on capability boundaries and trigger precision.
Evidence shows: README well-structured, install notes detailed, examples abundant, license clear, version number present but no changelog, maintenance responsibility not explicit. Deductions: missing known limitations and changelog.
Evidence shows: output usability (e.g., zip package) described, marginal value (e.g., local deployment) evident, but cost-benefit not quantified. Deductions: lack of cost-benefit data.
Evidence shows: claims in README partially supported by code, but independent verification lacking; cross-source corroboration limited; facts and inferences not clearly separated. Deductions: insufficient verification.
- Not found in source: confirmation before actingTurn on (or add) a confirmation step before it acts, and try it in a sandbox or test environment before real data.
- Not found in source: rollback or recovery pathBack up first, or work on a git branch or snapshot, so its changes can be undone.
- No concrete implementation of least privilege, user confirmation, rollback; assess before deployment.
- Dependency security not audited; check for vulnerabilities.
- Test coverage limited; not all features verified.