MCP and Skills

The desktop app supports the Model Context Protocol (MCP) for extensible tool integration and a skills system for adding specialized agent capabilities.

Model Context Protocol (MCP)

MCP is an open standard that lets AI agents interact with external tools and services. The desktop app includes built-in MCP servers and supports installing additional ones.

Built-in MCP Servers

These servers are included and automatically available:

Google Services Tools

When connected via OAuth, the Google MCP server provides tools across:

• Gmail -- Read, send, search, manage emails
• Calendar -- Create, update, list events
• Drive -- Upload, download, search files
• Photos -- Access and organize photos
• Meet -- Schedule and manage meetings
• Tasks -- Create and manage task lists
• Contacts -- Search and manage contacts
• Sheets -- Read and write spreadsheet data
• Slides -- Create and modify presentations
• Docs -- Read and edit documents

Tools are scope-filtered based on the permissions you grant during OAuth.

Installing MCP Servers

From the Marketplace

1. Go to Settings > MCP
2. Browse the Marketplace tab
3. Click Install on any server you want
4. The server is configured and ready to use

Available presets include:

• context7 -- Up-to-date library documentation
• sequential-thinking -- Enhanced reasoning chains
• filesystem -- File system operations
• github -- GitHub repository management
• playwright -- Browser automation
• slack -- Slack messaging

Remote OAuth Presets

Some MCP servers use Remote OAuth for secure authentication:

• Notion -- Access Notion workspaces
• Figma -- Design file access
• Granola -- Meeting notes

These use RFC 7591 Dynamic Client Registration with PKCE for secure connections.

Manual Configuration

MCP servers are configured from two sources:

• ~/.claude/settings.json -- Shared with Claude Code
• ~/.<app-slug>/mcp.json -- Application-specific servers

Supported transport protocols:

• stdio -- Local process communication
• HTTP -- Remote server via URL
• SSE -- Server-Sent Events
• Remote OAuth -- OAuth-authenticated remote servers

Skills System

Skills are reusable instruction sets that extend what the agent can do. They work like specialized "modes" the agent can activate.

Skill Format

Each skill is a SKILL.md file with YAML frontmatter:

---
name: my-skill
description: What this skill does
version: 1.0.0
author: username
---

Instructions for the agent when this skill is active...

Skill Sources

Skills are loaded from multiple directories:

Marketplace

The skills marketplace in Settings > Skills provides:

• Browse -- Paginated list with search
• Search -- Filter by name, slug, or author
• One-click install -- Install any skill instantly
• Update detection -- See when newer versions are available
• Version history -- View published dates and changelogs

Creating Skills

1. Go to Settings > Skills
2. Click Create Skill
3. Enter a name (automatically converted to kebab-case slug)
4. Write the skill instructions in the editor
5. Save -- the skill is immediately available to the agent

How Skills Work

When a skill is installed, its instructions are loaded into the agent's context. This lets the agent:

• Follow specialized workflows
• Use domain-specific knowledge
• Apply particular coding patterns or styles
• Interact with specific services in prescribed ways

Skills are non-destructive -- they add capabilities without removing existing ones.

Learn More

• Agent System -- How agents use tools during execution
• Desktop Application -- Overview and setup