Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →borski/award-calendar
Find the cheapest award dates for a route over a date range. Searches seats.aero across all programs, groups by date, and shows a calendar grid of the best deals.
Passed 13 of 19 test cases with the skill, and 8 of 19 without it.
and 3 more in the full report
6 fixed9 passed either way7 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 6 test cases the skill fixed. Same model, same prompt, with and without the skill.
When formatting a daily award calendar matrix for a date range where Sep 10 has zero business class award availability, what exact placeholder formatting should be used in the status column for that date?
| Checked | Without skill | With skill |
|---|---|---|
| Displays ❌ No availability in the status column for unavailable dates. | Missed | Passed |
The instructions your model reads when the skill loads.
"When should I fly SFO to NRT on points?" answered with a calendar view.
Searches seats.aero across a date range and presents a grid showing the cheapest award per day, per cabin. Highlights the best dates to fly.
This is an orchestration skill. It tells the agent how to query seats.aero and format the results. No standalone script.
seats-aero directly)Use the seats.aero cached search API with a date range. The API supports up to ~60 days per query.
curl -s -H "Partner-Authorization: $SEATS_AERO_API_KEY" \
"https://seats.aero/partnerapi/search?origin_airport={ORIGIN}&destination_airport={DEST}&start_date={START}&end_date={END}&cabin={CABIN}"
Parameters:
origin_airport: Can be comma-delimited for nearby airports (e.g., SFO,SJC,OAK)destination_airport: Same. Use multiple for metro areas (e.g., NRT,HND for Tokyo)start_date, end_date: YYYY-MM-DD. Keep range under 60 days for best results.cabin: Optional filter. economy, premium, business, first. Omit for all cabins.order_by: Use lowest_mileage to get cheapest first.take: Set to 1000 to get all results in one page.test cases passed, vs 8 / 19 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.