Skill: long-horizon-itinerary-sequencing

[Case 2]

• Initial Environment: An agent has access to a real-time database of 3.4 million flights, 3,892 restaurants, and 5,043 attractions with precise coordinates and public transit metadata (GTFS) for 140 U.S. cities.
• Real Question: Plan a 3-day trip for 1 person from El Paso to Seattle from November 1st to November 3rd, 2024, with a budget of $2,100. Include details of sports events. The traveler is an Adventure Seeker who likes cultural exploration and mountains.
• Real Trajectory: 1. Identify 'Seattle' as the destination and 'Adventure Seeker' as the persona. 2. Filter flights: Select Flight F0240857 (EP to Seattle, 18:47-21:25). 3. Search events: Identify 'Washington Huskies Football vs. USC Trojans' in Seattle. 4. Draft Day 1 POI list: Check-in at 'Cozy room near UW' and dine at 'Barolo Ristorante' from 22:00 to 23:00. 5. Ground transit: Verify the restaurant is 37.77m from 'Westlake And 7th' transit stop. 6. Sequence Day 2 around the sports event while maintaining 4-hour meal gaps.
• Real Answer: A granular itinerary where every POI includes a start/end timestamp and the distance to the nearest transit stop (e.g., visit Barolo Ristorante from 22:00 to 23:00, 37.77m from Westlake stop).
• Why this demonstrates the capability: This case demonstrates fine-grained spatio-temporal sequencing. The agent must not only find a valid path between cities but also schedule specific hours (22:00-23:00) and ground the plan in exact transit proximity (37.77m), ensuring the plan is spat-temp consistent down to the minute and meter.

Pipeline Execution Instructions

To synthesize data for this capability, you must strictly follow a 3-phase pipeline. Do not hallucinate steps. Read the corresponding reference file for each phase sequentially:

1. Phase 1: Environment Exploration Read the exploration guidelines to discover raw knowledge seeds: references/EXPLORATION.md

2. Phase 2: Trajectory Selection Once Phase 1 is complete, read the selection criteria to evaluate the trajectory: references/SELECTION.md

3. Phase 3: Data Synthesis Once a trajectory passes Phase 2, read the synthesis instructions to generate the final data: references/SYNTHESIS.md