Skill instructions
flightqueue-baggage1.33 KB
View saved version →
---
name: flightqueue-baggage
description: Look up airline baggage rules and lost-bag records with the FlightQueue MCP server. Use when the user asks "what is the carry-on size for Delta", "how much is a checked bag on United", "which airline loses the fewest bags", or "is it safe to check a bag on this airline".
---
# FlightQueue baggage
Connect the FlightQueue MCP server at `https://mcp.flightqueue.com/mcp`.
Both tools are free. The data covers US airlines and comes from US Bureau of
Transportation Statistics (BTS) reports.
## One airline
`get_airline_baggage_policy` with `carrier` — an airline slug such as
`american-airlines` or `delta`, or a 2-letter IATA code such as `AA`.
It returns carry-on size, checked-bag fees, the mishandled-bag rate and the
trend.
## All airlines
`get_airline_baggage_rankings` with optional `limit` (default 20, max 50)
— airlines ranked by how rarely they mishandle bags.
## Present the result
- For a policy question, give the carry-on size and the first checked-bag fee
first. Then give the lost-bag rate.
- For a ranking, show the top and bottom airlines with their rates, and the
period the BTS data covers.
- Fees change often. Tell the user to confirm the fee on the airline's own site
before they pay.
- A non-US airline returns "not found". Say that the dataset is US-only; do not
guess a policy.
flightqueue-compare-airports1.66 KB
View saved version →
---
name: flightqueue-compare-airports
description: Choose which airport to fly from with the FlightQueue MCP server. Use when the user asks "should I fly from JFK or Newark", "which London airport has the shortest security queue", "compare LAX and SFO", or which hub to connect through.
---
# FlightQueue compare airports
Connect the FlightQueue MCP server at `https://mcp.flightqueue.com/mcp`.
## Workflow
1. `search_airports` with `query` (a city or name) and optional `country`
(ISO-2). Use it to turn "London" or "New York" into real airport codes.
Metro codes such as LON and NYC are not airports.
2. `compare_airports` with `iata_codes` — a list of 2 to 5 codes. It returns
average security and immigration waits side by side.
3. `get_airport_delays_summary` with `iata` for each candidate — current FAA
ground stops and delay programmes. Major US airports only; other airports
return an `unknown` status. Do not read `unknown` as "no delays".
4. `get_historical_wait_times` with `iata` **(paid)** — daily and weekly
patterns. Use it when the user travels on a known day and time, e.g. "Friday
evening".
## Present the result
Show a small table: airport, average security wait, immigration wait (if
international), current delay status. Then give one recommendation and the
main reason for it.
Averages do not show today's queue. For a flight today, add the current wait
from the `flightqueue-trip-day` skill.
## Paid tools
`get_historical_wait_times` needs a FlightQueue Premium plan. When it
returns a Premium error, do not retry. Give the comparison from the free tools
and tell the user that day-by-day patterns are part of
https://flightqueue.com/pricing.
flightqueue-crowd-reports1.79 KB
View saved version →
---
name: flightqueue-crowd-reports
description: Read traveller-submitted airport wait reports, and submit one when the user is at the airport, with the FlightQueue MCP server. Use when the user asks "what are people reporting at security", "is the queue really that long", or says "I just got through security in 20 minutes".
---
# FlightQueue crowd reports
Connect the FlightQueue MCP server at `https://mcp.flightqueue.com/mcp`.
Both tools are free and need no account.
## Read reports
`get_crowd_reports` with `iata` and optional `checkpoint_type`:
`security`, `immigration`, `checkin` or `all` (default).
Each report has a terminal, a time and a confidence. Reports are a small,
noisy sample. Use them to check a predicted wait, not to replace it. Say how
many reports there are and how old the newest one is.
## Submit a report
`submit_wait_time_report` writes to the shared prediction model. A wrong
report makes predictions worse for every traveller.
Call it only when **all** of these are true:
1. The user is at the airport now, or has just passed the queue.
2. The user gave the wait they actually had.
3. The user agreed to share it. Ask first: "Do you want to share this wait
with other travellers?"
Never submit a number you read from another tool, a guess, or an estimate.
Arguments:
| Field | Required | Notes |
|---|---|---|
| `iata` | yes | 3-letter airport code |
| `wait_time_minutes` | yes | 0-180 |
| `checkpoint_type` | no | `security` (default), `immigration`, `checkin`, `customs`, `baggage`, `lounge`, `ees_kiosk` |
| `terminal` | no | e.g. `4` or `B` |
| `precheck` | no | `true` for a TSA PreCheck or fast-track lane |
| `at_airport_now` | no | `true` when the user is at the airport |
Set `at_airport_now: true` when it applies. After the call, tell the user the
report was recorded.
flightqueue-ees-schengen2.01 KB
View saved version →
---
name: flightqueue-ees-schengen
description: Check EU Entry/Exit System (EES) border queues and the Schengen 90/180-day allowance with the FlightQueue MCP server. Use when the user asks "how long is passport control at Lisbon", "what is the EES queue", "how many Schengen days do I have left", or "will this trip overstay".
---
# FlightQueue EES and Schengen
Connect the FlightQueue MCP server at `https://mcp.flightqueue.com/mcp`.
All three tools are free.
## Border queue now
`get_ees_wait_times` returns first-entry (biometric registration) and
returning-traveller waits per airport.
- One airport: pass `iata`.
- One country: pass `country` as an ISO-2 code, e.g. `PT`.
- The full board: pass neither. `limit` caps the list (default 25, max 150).
Each row has a confidence tier. Say whether a number is a live reading or a
modelled estimate.
## Border queue later
`get_ees_forecast` with `iata` gives a 24-hour forecast. Set `direction`
to `entry` for an arriving traveller, `exit` for a departing one, or
`both` (default).
A first-time EES traveller registers fingerprints and a photo at the border.
Use the first-entry figure for them, not the returning-traveller one.
## Schengen 90/180
`check_schengen_days` counts days used in any 180-day window.
- `trips`: a list of `{ entryDate, exitDate?, country? }` with ISO dates
(`YYYY-MM-DD`). Leave out `exitDate` for a stay that is still in progress.
- To test a future trip, add `planned_entry_date` and `planned_days`.
Entry and exit days both count, and overlapping trips count once. The tool
runs offline, so the trip history does not leave the tool.
Ask the user for their trip dates. Never invent dates to fill a gap. If the
result is close to 90 days, say so clearly and tell the user to confirm with
the official EU calculator before they travel. This is not legal advice.
## Present the result
Give the wait in minutes as a range, the airport, the direction, and the
local time of the reading. For Schengen, give days used, days left, and the
date the next day becomes free.
flightqueue-trip-day2.2 KB
View saved version →
---
name: flightqueue-trip-day
description: Plan the day of a flight with the FlightQueue MCP server. Use when the user asks "when should I leave for the airport", "what time should I get to JFK", "is my airport delayed today", or which terminal or lounge to use.
---
# FlightQueue trip day
Connect the FlightQueue MCP server at `https://mcp.flightqueue.com/mcp`.
Every tool below takes `iata`, a 3-letter airport code. If the user gives a
city or airport name, call `search_airports` with `query` first. Never guess
the code.
## Workflow
1. `get_best_time_to_fly` — the quietest arrival window, from historical
patterns. Free keys get one recommendation; paid keys get the top 3 with
checkpoint detail.
2. The current queue:
- `get_live_wait_time` **(paid)** — the freshest reading, from the
airport's own feed. Try it first when the user has Premium.
- `get_security_wait_estimate` — the free snapshot. Use it when the live
call fails with a Premium error, or when the user has no plan.
3. `get_faa_delays` with `airports_only_with_delays: true` — ground stops
and delay programmes. US airports only.
4. `get_aviation_weather` — METAR, TAF and flight category. Use it to explain
*why* a delay exists: low cloud, visibility, wind or storms.
5. `get_terminals` with `airline` — which terminal and checkpoint the user's
airline uses.
6. `get_lounges` with `terminal` — lounges airside of that checkpoint, with
access rules.
Skip the steps the user did not ask about. A question about the queue alone
needs steps 1 and 2.
## Present the result
Lead with one time: "Leave to reach the airport by 06:10." Then give the
reasons in this order: security queue, delays, weather, terminal, lounge.
- Say which source each wait came from. A live reading and a free estimate are
different things; never present an estimate as observed.
- Quote a range, not a single minute value, when the tool returns a confidence
band.
- All times are in the airport's local time.
## Paid tools
`get_live_wait_time` needs a FlightQueue Premium plan. When it returns a
Premium error, do not retry. Fall back to `get_security_wait_estimate` and
tell the user that live waits are part of https://flightqueue.com/pricing.