What problem does it solve? Generic AI travel advice often returns unordered lists of attractions, stale opening hours, and vague directions. This Skill turns natural-language travel requests into executable plans by parsing city, duration, crowd preference, budget, and travel type, then routing the request to recommendation rules, weather lookups, Amap route planning, or web search for the latest ticket and reservation policies. ## Core Features & Use Cases - Scenario-Based Recommendations: Filters attractions and restaurants by crowd level (S/A/B/C/D), duration (half-day/full-day/evening), budget, and travel type (couple, friends, family, solo), then outputs an executable play-plus-eat route instead of a raw list. - Tool-Augmented Queries: Calls QWeather for current conditions and 3-day forecasts with travel advice, Amap for geocoding and walking/driving/transit routes, and DuckDuckGo for the latest opening hours and booking policies. - Version Snapshots: Built-in snapshot and rollback tooling lets you version SKILL.md, references, and scripts before major edits. - Use Case: Ask "Plan a half-day in Hangzhou for a couple, avoid crowds, include dinner" and receive a structured route with a quiet West Lake walking segment, a tea-house stop, and a local restaurant with price and queue notes. ## Quick Start Ask the AI to plan a half-day couple-friendly itinerary in Hangzhou with low crowds and a local dinner included.
