Run2 itinerary optimization

Skill cxcscmu/SkillLearnBench/skills/b2-self-feedback-claude-haiku-4-5/travel-planning/run2_itinerary_optimization

[COLM'26] SkillLearnBench is the first benchmark for evaluating continual learning methods that automatically generate agent skills.

Install

npx -y skills add cxcscmu/SkillLearnBench --skill run2_itinerary_optimization

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

What its author says it does

Copied from the file, not written here

Optimized itinerary building with cuisine distribution, cost tracking, and multi-city routing for pet-friendly travel.

SKILL.md

4.8 KB, ~1.2k tokens by cl100k_base, as published. Nobody here has run it

Optimized Multi-City Itinerary Planning

Overview

This skill addresses:

Balanced cuisine distribution across 7 days
Cost tracking and budget adherence
Optimal city sequencing and travel timing
Pet-friendly accommodation prioritization
Realistic meal and attraction planning

Route Optimization for 3 Ohio Cities

Recommended Route

Minneapolis → Cleveland → Columbus → Cincinnati → Minneapolis

Rationale:

Cleveland is north, good entry point
Columbus is central, on route to Cincinnati
Cincinnati is southwest, good final city before returning
Minimizes backtracking

Day Allocation

Day 1: Travel Minneapolis → Cleveland (arrival dinner)
Days 2-3: Cleveland (full exploration)
Day 4: Travel Cleveland → Columbus (arrival dinner)
Day 5: Columbus (full exploration)
Day 6: Travel Columbus → Cincinnati (arrival dinner)
Day 7: Cincinnati & return to Minneapolis (depart dinner/lunch only)

Cuisine Distribution Strategy

Assignment Pattern

Across 7 days, ensure all 4 cuisines appear:

American: 2-3 occurrences (breakfast Day 1, lunch Day 5, dinner Day 7)
Italian: 2 occurrences (dinner Days 2, 6)
Chinese: 2 occurrences (lunch Days 3, 7)
Mediterranean: 1 occurrence (lunch Day 4)

Implementation

cuisine_schedule = {
    1: {"breakfast": "American", "lunch": "American", "dinner": "Chinese"},
    2: {"breakfast": "Italian", "lunch": "Mediterranean", "dinner": "Italian"},
    3: {"breakfast": "Chinese", "lunch": "American", "dinner": "American"},
    4: {"breakfast": "Mediterranean", "lunch": "Mediterranean", "dinner": "Italian"},
    5: {"breakfast": "American", "lunch": "Italian", "dinner": "Chinese"},
    6: {"breakfast": "Italian", "lunch": "Chinese", "dinner": "Italian"},
    7: {"breakfast": "American", "lunch": "Chinese", "dinner": "-"}
}

Cost Tracking

Budget Breakdown ($5,100 for 2 people, 7 days)

Accommodations: 6 nights × $120-180/night = $720-1,080
Meals: 7 days × $80-120/day = $560-840
Attractions: $200-300 total (entry fees)
Gas/Transportation: $150-200 (est. 1,000 miles round trip)
Buffer: $200-500

Daily Budgets

daily_budget = {
    "accommodation_per_night": 150,  # per night for 2 people
    "meals_per_day": 100,  # ~$17 per person per meal
    "attractions_per_day": 30-50,
    "transport_per_day": 20-30
}

Meal Planning Constraints

No Restaurants Repeated Across Days
- Even if good, pick different one next time
- Keep cuisine diversity high
Format Rule
- Include restaurant name + cuisine type + city
- Example: "Italian at Via Cento, Cleveland"
- Never: Just restaurant name or just cuisine name
Travel Days
- Breakfast: lighter meal in departure city
- Lunch: en route or arrival city
- Dinner: arrival city (establish evening routine)

Attraction Selection

Strategy

2-3 attractions per day minimum
Mix types: museums, parks, historic sites, food markets
Prioritize based on:
1. Availability in city
2. Pet-friendly (where applicable - parks over indoor museums)
3. Rating/popularity
4. Distance from accommodation

Format

List attractions separated by semicolon
Must end with semicolon: "Attraction1;Attraction2;"
Never end without semicolon or without enough attractions

Accommodation Selection Priority

Pet-friendly verification (explicitly no "No pets")
Price within budget ($100-200/night)
Rating > 4.0
Room type: Entire home/apt > Private room > Shared room
Minimum nights requirement = 1

Naming Convention

Include: "Pet-friendly" + accommodation type + city name
Example: "Pet-friendly Studio Apartment, Cleveland"
Avoid: Excessive special characters, Unicode issues

Travel Day Optimization

def plan_travel_day(origin_city, dest_city):
    """Structure for travel days"""
    return {
        "current_city": f"from {origin_city} to {dest_city}",
        "transportation": f"Self-driving: from {origin_city} to {dest_city}",
        "breakfast": "meal in origin city",
        "lunch": "meal en route or early arrival",
        "dinner": "meal in destination city",
        "attraction": "light attractions in destination",
        "accommodation": "in destination city"
    }

Quality Assurance Checklist

Before finalizing each day:

Exactly one city or "from X to Y" format
Transportation includes direction and mode
All meal fields filled (no blank values)
Cuisines match assigned schedule
Attractions string ends with semicolon
Accommodation marked pet-friendly
No restaurant repeats within 2 days
Budget for day stays within limits