google-maps
Installation
SKILL.md
Google Maps - Geospatial Query Capabilities
Overview
Gives an AI Agent the ability to reason about physical space — understand locations, distances, routes, and elevation, and naturally weave that information into conversation.
Without this Skill, the agent can only guess or refuse when asked "how do I get from Taipei 101 to the National Palace Museum?". With it, the agent returns exact coordinates, step-by-step routes, and travel times.
Core Principles
| Principle | Explanation |
|---|---|
| Chain over single-shot | Most geo questions require 2-5 tool calls chained together. See Scenario Recipes in references/tools-api.md for the full patterns. |
| Match recipe to intent | Map the user's question to a recipe (Trip Planning, Local Discovery, Route Comparison, Neighborhood Analysis, Multi-Stop, Place Comparison, Along the Route) before calling any tool. |
| Precise input saves trouble | Use coordinates over address strings when available. Use place_id over name search. More precise input = more reliable output. |
| Output is structured | Every tool returns JSON. Use it directly for downstream computation or comparison — no extra parsing needed. |
| Present as tables | Users prefer comparison tables and scorecards over raw JSON. Format results for readability. |