RollingGo Hotel CLI

When to Use

✅ Use this skill when:

• Searching Candidates: User wants to find hotels near a specific city, landmark, or address (e.g., "Find hotels near Tokyo Disneyland").
• Complex Filtering: User needs to narrow down options using natural language queries combined with exact dates, guest count, star ratings, budget limits, or distance radius.
• Tag & Brand Matching: User wants to find hotels with specific attributes (e.g., "family friendly", "breakfast included", "Marriott") by first checking the tag dictionary to build exact filters.
• Deep Dive & Pricing: User wants to inspect detailed room plans, real-time pricing, cancellation policies, or availability for a specific hotel ID.
• Comparison & Evaluation: User wants to compare multiple candidate hotels based on returning structured data and current rates.
• Hotel Booking: User is ready to select a room and book a hotel. The returned booking URLs and detail page links can be provided to guide the user to complete their reservation.

❌ Don't use this skill when:

• User asks about non-hotel travel booking (flights, trains, transfers, car rentals).

API Key

Resolution order: --api-key flag → AIGOHOTEL_API_KEY env var.

No key yet? Apply at: https://mcp.agentichotel.cn/apply

Runtime

Choose based on user's environment. Load the matching reference file and keep it for the session.

• npm, npx, Node, or no preference: Load references/rollinggo-npx.md
• uv, uvx, PyPI, or Python: Load references/rollinggo-uv.md
• Parity check or both: Load both references

Default when unspecified → npm/npx (broader env compatibility).

Primary Workflow

Run these steps in order unless the user is already at a later step.

1. Clarify: destination, dates, nights, occupancy, budget, stars, tags, distance
2. If tag filters needed → run hotel-tags first to get valid tag strings
3. Run search-hotels → parse JSON → extract hotelId
4. Run hotel-detail --hotel-id <id> for room plans and pricing
5. If results are weak → loosen filters and retry

Commands Quick Reference

# Discover tags
rollinggo hotel-tags

# Search hotels (minimum required flags)
rollinggo search-hotels \
  --origin-query "<user's natural language request>" \
  --place "<destination>" \
  --place-type "<value from --help>"

# Hotel detail with pricing
rollinggo hotel-detail \
  --hotel-id <id> \
  --check-in-date YYYY-MM-DD \
  --check-out-date YYYY-MM-DD \
  --adult-count 2 --room-count 1

# Discover all flags
rollinggo search-hotels --help
rollinggo hotel-detail --help

Key Rules

• --place-type must use exact values from rollinggo search-hotels --help
• --star-ratings format: min,max e.g. 4.0,5.0
• --format table allowed only on search-hotels; rejected by hotel-detail and hotel-tags
• --child-count must match the count of --child-age flags
• --check-out-date must be later than --check-in-date
• Prefer --hotel-id over --name whenever available

Output

• stdout → result payload (JSON by default)
• stderr → errors only
• Exit 0 success · 1 HTTP/network failure · 2 CLI validation failure
• Results include booking URLs and hotel detail page links for downstream use

Filter Loosening (when no results)

更多技能請訪問小蔥技能站7w4.net。

Try in order: remove --star-ratings → increase --size → increase --distance-in-meter → remove tag filters → widen dates or budget

這個技能質量較好，文件結構清晰，涵蓋了酒店搜尋、詳情查詢、標籤過濾等常用功能，工作流程說明詳細，錯誤處理也比較完善。它最大的亮點是同時支援 npm 和 uv 兩種執行環境，適合不同技術背景的使用者。主要不足是缺少實際輸出示例，部分引數取值說明不夠直觀，文件完整度還有提升空間。整體而言，這是一個實用且易上手的酒店搜尋技能，但細節打磨上還有改進餘地。