Update to Mapbox

{
  "name": "mapbox-mcp-runtime-patterns",
  "description": "Integration patterns for Mapbox MCP Server in AI applications and agent frameworks. Covers runtime integration with pydantic-ai, mastra, LangChain, and custom agents. Use when building AI-powered applications that need geospatial capabilities.",
  "included_files": [
    {
      "relative_path": "AGENTS.md",
      "size_in_bytes": 12928
    },
    {
      "relative_path": "evals/evals.json",
      "size_in_bytes": 4008
    },
    {
      "relative_path": "examples/README.md",
      "size_in_bytes": 7409
    },
    {
      "relative_path": "examples/python/crewai_example.py",
      "size_in_bytes": 9319
    },
    {
      "relative_path": "examples/python/pydantic_ai_example.py",
      "size_in_bytes": 5694
    },
    {
      "relative_path": "examples/python/requirements.txt",
      "size_in_bytes": 220
    },
    {
      "relative_path": "examples/python/smolagents_example.py",
      "size_in_bytes": 9417
    },
    {
      "relative_path": "examples/typescript/langchain-example.ts",
      "size_in_bytes": 7896
    },
    {
      "relative_path": "examples/typescript/mastra-example.ts",
      "size_in_bytes": 8112
    },
    {
      "relative_path": "examples/typescript/package-lock.json",
      "size_in_bytes": 119371
    },
    {
      "relative_path": "examples/typescript/package.json",
      "size_in_bytes": 654
    },
    {
      "relative_path": "examples/typescript/tsconfig.json",
      "size_in_bytes": 393
    },
    {
      "relative_path": "references/crewai.md",
      "size_in_bytes": 7086
    },
    {
      "relative_path": "references/custom-agent.md",
      "size_in_bytes": 7599
    },
    {
      "relative_path": "references/langchain.md",
      "size_in_bytes": 5861
    },
    {
      "relative_path": "references/mastra.md",
      "size_in_bytes": 2667
    },
    {
      "relative_path": "references/production.md",
      "size_in_bytes": 9778
    },
    {
      "relative_path": "references/pydantic-ai.md",
      "size_in_bytes": 2903
    },
    {
      "relative_path": "references/smolagents.md",
      "size_in_bytes": 6186
    },
    {
      "relative_path": "references/use-cases.md",
      "size_in_bytes": 2980
    }
  ],
  "skill_md_contents": "---\nname: mapbox-mcp-runtime-patterns\ndescription: Integration patterns for Mapbox MCP Server in AI applications and agent frameworks. Covers runtime integration with pydantic-ai, mastra, LangChain, and custom agents. Use when building AI-powered applications that need geospatial capabilities.\n---\n\n# Mapbox MCP Runtime Patterns\n\nThis skill provides patterns for integrating the Mapbox MCP Server into AI applications for production use with geospatial capabilities.\n\n## What is Mapbox MCP Server?\n\nThe [Mapbox MCP Server](https://github.com/mapbox/mcp-server) is a Model Context Protocol (MCP) server that provides AI agents with geospatial tools:\n\n**Offline Tools (Turf.js):**\n\n- Distance, bearing, midpoint calculations\n- Point-in-polygon tests\n- Area, buffer, centroid operations\n- Bounding box, geometry simplification\n- No API calls, instant results\n\n**Mapbox API Tools:**\n\n- Directions and routing\n- Reverse geocoding\n- POI category search\n- Isochrones (reachability)\n- Travel time matrices\n- Static map images\n- GPS trace map matching\n- Multi-stop route optimization\n\n**Utility Tools:**\n\n- Server version info\n- POI category list\n\n**Key benefit:** Give your AI application geospatial superpowers without manually integrating multiple APIs.\n\n## Understanding Tool Categories\n\nBefore integrating, understand the key distinctions between tools to help your LLM choose correctly:\n\n### Distance: \"As the Crow Flies\" vs \"Along Roads\"\n\n**Straight-line distance** (offline, instant):\n\n- Tools: `distance_tool`, `bearing_tool`, `midpoint_tool`\n- Use for: Proximity checks, \"how far away is X?\", comparing distances\n- Example: \"Is this restaurant within 2 miles?\" → `distance_tool`\n\n**Route distance** (API, traffic-aware):\n\n- Tools: `directions_tool`, `matrix_tool`\n- Use for: Navigation, drive time, \"how long to drive?\"\n- Example: \"How long to drive there?\" → `directions_tool`\n\n### Search: Type vs Specific Place\n\n**Category/type search**:\n\n- Tool: `category_search_tool`\n- Use for: \"Find coffee shops\", \"restaurants nearby\", browsing by type\n- Example: \"What hotels are near me?\" → `category_search_tool`\n\n**Specific place/address**:\n\n- Tool: `search_and_geocode_tool`, `reverse_geocode_tool`\n- Use for: Named places, street addresses, landmarks\n- Example: \"Find 123 Main Street\" → `search_and_geocode_tool`\n\n### Travel Time: Area vs Route\n\n**Reachable area** (what's within reach):\n\n- Tool: `isochrone_tool`\n- Returns: GeoJSON polygon of everywhere reachable\n- Example: \"What can I reach in 15 minutes?\" → `isochrone_tool`\n\n**Specific route** (how to get there):\n\n- Tool: `directions_tool`\n- Returns: Turn-by-turn directions to one destination\n- Example: \"How do I get to the airport?\" → `directions_tool`\n\n### Cost & Performance\n\n**Offline tools** (free, instant):\n\n- No API calls, no token usage\n- Use whenever real-time data not needed\n- Examples: `distance_tool`, `point_in_polygon_tool`, `area_tool`\n\n**API tools** (requires token, counts against usage):\n\n- Real-time traffic, live POI data, current conditions\n- Use when accuracy and freshness matter\n- Examples: `directions_tool`, `category_search_tool`, `isochrone_tool`\n\n**Best practice:** Prefer offline tools when possible, use API tools when you need real-time data or routing.\n\n## Installation & Setup\n\n### Option 1: Hosted Server (Recommended)\n\n**Easiest integration** - Use Mapbox's hosted MCP server at:\n\n```\nhttps://mcp.mapbox.com/mcp\n```\n\nNo installation required. Simply pass your Mapbox access token in the `Authorization` header.\n\n**Benefits:**\n\n- No server management\n- Always up-to-date\n- Production-ready\n- Lower latency (Mapbox infrastructure)\n\n**Authentication:**\n\nUse token-based authentication (standard for programmatic access):\n\n```\nAuthorization: Bearer your_mapbox_token\n```\n\n**Note:** The hosted server also supports OAuth, but that's primarily for interactive flows (coding assistants, not production apps).\n\n### Option 2: Self-Hosted\n\nFor custom deployments or development:\n\n```bash\nnpm install @mapbox/mcp-server\n```\n\nOr use directly via npx:\n\n```bash\nnpx @mapbox/mcp-server\n```\n\n**Environment setup:**\n\n```bash\nexport MAPBOX_ACCESS_TOKEN=\"your_token_here\"\n```\n\n## Reference Files\n\nDetailed integration patterns and production guidance are organized into reference files. Load the ones relevant to your task.\n\n- **Pydantic AI** -- Type-safe Python agents\n  Load: `references/pydantic-ai.md`\n\n- **CrewAI** -- Multi-agent orchestration\n  Load: `references/crewai.md`\n\n- **Smolagents** -- Lightweight HuggingFace agents\n  Load: `references/smolagents.md`\n\n- **Mastra** -- Multi-agent TypeScript systems\n  Load: `references/mastra.md`\n\n- **LangChain** -- Conversational AI with tool chaining\n  Load: `references/langchain.md`\n\n- **Custom Agent** -- Zillow/TripAdvisor/DoorDash-style patterns, architecture diagrams, hybrid approach\n  Load: `references/custom-agent.md`\n\n- **Use Cases** -- Real Estate, Food Delivery, Travel Planning examples\n  Load: `references/use-cases.md`\n\n- **Production Patterns** -- Caching, batch operations, tool descriptions, error handling, security, rate limiting, testing\n  Load: `references/production.md`\n\n## Resources\n\n- [Mapbox MCP Server](https://github.com/mapbox/mcp-server)\n- [Model Context Protocol](https://modelcontextprotocol.io)\n- [Pydantic AI](https://ai.pydantic.dev/)\n- [Mastra](https://mastra.ai/)\n- [LangChain](https://docs.langchain.com/oss/javascript/langchain/overview/)\n- [Mapbox API Documentation](https://docs.mapbox.com/api/)\n\n## When to Use This Skill\n\nInvoke this skill when:\n\n- Integrating Mapbox MCP Server into AI applications\n- Building AI agents with geospatial capabilities\n- Architecting Zillow/TripAdvisor/DoorDash-style apps with AI\n- Choosing between MCP, direct APIs, or SDKs\n- Optimizing geospatial operations in production\n- Implementing error handling for geospatial AI features\n- Testing AI applications with geospatial tools\n"
}