tasi-travel

Generate day-by-day travel itineraries from user constraints with provider-backed recommendations.

9|2|Updated Apr 24, 2026
One-click install
npx skills add https://github.com/TasiTech/tasi-harness --skill tasi-travel
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: tasi-travel
Source: https://github.com/TasiTech/tasi-harness/tree/main/resources/skills/travel/tasi-travel
Command: npx skills add https://github.com/TasiTech/tasi-harness --skill tasi-travel

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) and assets (resource) and scripts (resource) components.

What problem does it solve?

It turns complex travel planning constraints into a structured, day-by-day itinerary with provider-backed recommendations and clear uncertainty when data quality degrades.

Core Features & Use Cases

  • Constraint-driven trip planning: Accepts destination, date range, traveler count, budget, and optional must-visit/avoid/hotel/transport preferences to produce an actionable schedule.
  • Provider routing with graceful fallback: Uses a capability-to-provider priority table (including browser-first Ctrip retrieval) and switches providers when results are missing, degraded, or blocked.
  • Evidence and traceability: Preserves source URLs (and item-level detail URLs when available) so itinerary items can be cited, while explicitly labeling inferred/estimated values.
  • Optional Amap route links: Generates per-day Amap driving/walking-style route previews when mapping is requested, respecting via-point and point-count limits.
  • Budgeting and verification guardrails: Outputs budget breakdown, transportation/hotel/dining recommendations, and runs a consistency verification pass (time logic, commute load, activity fit).

Quick Start

Ask tasi-travel to create a 5-day itinerary for Beijing for 2 travelers within a total budget of 8000 CNY, including must-visit items for Tiananmen Square and the Summer Palace, and generate Amap route links for each day.

Frequently Asked Questions about tasi-travel

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate a travel itinerary with evidence-backed recommendations?

To generate a travel itinerary with evidence-backed recommendations, provide your destination, dates, traveler count, budget, and preferences. The system uses browser-first Ctrip extraction to source hotels, flights, and POIs, preserving source URLs for citations while labeling estimated values.

Can I create route maps for a daily travel schedule automatically?

You can create route maps for a daily travel schedule by requesting Amap route links during itinerary generation. The system renders driving or walking route previews for each day, automatically respecting via-point and point-count limits.

How does travel planning handle missing or degraded hotel and flight data?

Travel planning handles missing or degraded hotel and flight data through a provider routing policy with graceful fallback. If browser-first Ctrip retrieval is blocked or returns poor results, it automatically switches to alternative providers to maintain data quality.

What inputs do I need for constraint-driven trip planning?

Constraint-driven trip planning requires YAML-defined inputs including destination, date range, traveler count, and budget. You can also include optional preferences like must-visit locations, places to avoid, hotel choices, and transportation methods.

Does the itinerary generation process verify budget and time logic?

Itinerary generation verifies budget and time logic through a consistency verification pass. It checks time logic, commute load, and activity fit, while outputting a detailed budget breakdown for transportation, hotels, and dining.

Why are some values in my generated itinerary labeled as uncertain?

Values in your generated itinerary are labeled as uncertain when data quality degrades or specific details are inferred rather than directly sourced. The system explicitly labels these estimated values to distinguish them from provider-backed, cited recommendations.