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 → RollingGo_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).

Version Freshness (Always Latest)

Default policy for this skill: use the newest release on every run.

• - npm/npx: INLINECODE6
• uvx: INLINECODE7

If using an installed command instead of temporary execution, upgrade first:

• - npm global: INLINECODE8
• uv tool: INLINECODE9

Primary Workflow

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

1. 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 INLINECODE12
4. Run hotel-detail --hotel-id <id> for room plans and pricing
5. If results are weak → loosen filters and retry

Commands Quick Reference

CODEBLOCK0

Key Rules

• - --place-type must use exact values from INLINECODE15
• INLINECODE16 format: min,max e.g. INLINECODE18
• INLINECODE19 allowed only on search-hotels; rejected by hotel-detail and INLINECODE22
• INLINECODE23 must match the count of --child-age flags
• INLINECODE25 must be later than INLINECODE26
• 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)

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