> ## Documentation Index > Fetch the complete documentation index at: https://openworklabs.com/docs/llms.txt > Use this file to discover all available pages before exploring further. > ## Agent Instructions > For agents: install the OpenWork desktop app with `brew install --cask openwork` (macOS) or from https://openworklabs.com/download. Do not run `npx openwork`; the npm package named `openwork` is a different project. > Connect an agent to OpenWork with the MCP Gateway: `claude mcp add --transport http openwork https://api.openworklabs.com/mcp/agent`. Teams sign up at https://app.openworklabs.com?mode=sign-up. > Step-by-step agent guide: https://openworklabs.com/docs/start-here/use-openwork-from-an-ai-agent.md > Connect Slack (any organization admin): https://openworklabs.com/docs/slack/connect-slack.md. OpenWork in Slack (@openwork bot) is private alpha for approved organizations. Setup guide: https://openworklabs.com/docs/slack/overview.md. Agent skill: https://openworklabs.com/.well-known/agent-skills/set-up-openwork-slack/SKILL.md. # Share skills and MCP servers with your team > Give teammates the same skills, plugins, connectors, and models through OpenWork Cloud, without copying files or editing MCP config. Set something up once in OpenWork Cloud, choose who can use it, and your teammates get it in the desktop app. Nobody copies folders or edits MCP config files. You can share three kinds of things: * **Skills and plugins**: instructions and bundles your team's AI follows. * **Connectors**: apps and MCP servers the AI can work in, like Slack or Notion. * **Models**: the AI models your team is allowed to use. > [!NOTE] > > **Local or managed.** Local workspace skills are free and need no OpenWork Cloud sign-in. You can copy their files to a teammate yourself. Sharing with access control and automatic updates uses OpenWork Cloud (or a self-hosted deployment) and is part of the Team and Enterprise plans. See [Pricing and licensing](https://openworklabs.com/docs/start-here/pricing-and-licensing.md). ## Share skills A skill is a set of plain-language instructions. Each skill lives in a plugin, and a plugin can hold one skill or several. 1. Create the skill or plugin in OpenWork Cloud. Any member can do this from **My Library**. 2. Open it and choose **Who can use it**: everyone in the organization, selected teams, or selected people. 3. Teammates find it in **Settings › Library** in the desktop app. Any member can share what they made with people and teams. Only admins can share with everyone in the organization. For the full steps, and for making your own copy of a shared skill, see [Publish and copy a skill](https://openworklabs.com/docs/start-here/do-work-with-it/publish-and-copy-a-skill.md). To give the same audience several plugins at once, an admin can group them in a [Collection](https://openworklabs.com/docs/cloud/share-with-your-team/collections.md). ## Share MCP servers Admins add MCP servers as connectors. Members never see server URLs or tokens. 1. In the OpenWork Cloud dashboard, open **Manage › Connectors** and click **Add connector**. 2. Pick a preset, like GitHub or Notion. To use another server, paste its URL in the filter box or click **Add any MCP**. 3. OpenWork checks the server. If it asks, sign in to the service or enter its API key and click **Save key**. 4. Under **How people sign in**, choose **Each person signs in** (everyone uses their own account) or **One account for everyone** (you sign in once for the team). 5. Under **Who can use it**, choose who gets it, then click **Add** followed by the connector name. 6. Members open **Settings › Library** in the desktop app. If each person signs in, they click **Sign in** on the connector once. For sign-in options, Slack, and troubleshooting, see [Sharing MCP connections](https://openworklabs.com/docs/cloud/share-with-your-team/shared-mcp-connections.md) and [MCP server compatibility](https://openworklabs.com/docs/cloud/share-with-your-team/mcp-server-compatibility.md). ## Use the same setup in other AI tools The same shared skills and connectors work outside the desktop app through the OpenWork MCP Gateway. [Add it to Claude Code, Codex, Cursor, VS Code, or another MCP client in one command](https://openworklabs.com/docs/start-here/connect-openwork-mcp.md). ## Share models too To give the team the same models, an admin adds a [managed](https://openworklabs.com/docs/cloud/share-with-your-team/managed-llm-provider.md) or [custom LLM provider](https://openworklabs.com/docs/cloud/share-with-your-team/custom-llm-provider.md) once. To control which settings members can change in the app, see [desktop policies](https://openworklabs.com/docs/cloud/share-with-your-team/desktop-policies.md). ## Share without OpenWork Cloud Skills are plain `SKILL.md` folders. A workspace keeps its skills in `.opencode/skills/`. You can copy a skill folder into a teammate's workspace or commit it to a shared repository. You lose access control and automatic updates, and each person still connects MCP servers on their own. ## FAQ > [accordion] Do my teammates need to be technical? > > No. Once something is shared with them, it shows up in the desktop app. At most, a member signs in to a service once. > [accordion] Who can publish skills for other people? > > Any member can create a skill or plugin and share it with selected people or teams. Only admins (owners, super-admins, and admins) can share with everyone in the organization, change sharing on plugins other people made, or manage Collections. > [accordion] If I update a shared skill, does everyone get the change? > > Yes. Everyone who has the skill gets the new version in their next task. To try a change on your own first, make a copy. See [Copy without changing the shared skill](https://openworklabs.com/docs/start-here/do-work-with-it/publish-and-copy-a-skill.md#copy-without-changing-the-shared-skill). > [accordion] Can I share a local (stdio) MCP server? > > No. Connectors are remote MCP servers reached by URL. For a server that only runs on your computer, add it on each machine with [Add a custom MCP server](https://openworklabs.com/docs/start-here/connect-your-stack/add-an-mcp-server.md). > [accordion] Can I run this on my own infrastructure? > > Yes. The organization control plane that manages plugins, connectors, and access can be [self-hosted](https://openworklabs.com/docs/start-here/self-host.md). ## Related * [Connectors, skills, and plugins](https://openworklabs.com/docs/start-here/do-work-with-it/skills-plugins-and-mcp.md) * [Team quickstart](https://openworklabs.com/docs/cloud/team-quickstart.md) * [Migrate from Claude Cowork](https://openworklabs.com/docs/start-here/migrate-from-claude-cowork.md) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.