agentleFS
mtnrabimtnrabi/travel-agent-skills/skills/trip-planner/SKILL.md

trip-planner

Skill4 starsChanged 23 days agoMITView on GitHub

Plan and price a whole trip for a named destination by combining a real-time round-trip Google Flights search with a Booking.com hotel search over the same dates, aligning checkin_date and checkout_date to the flight dates, deriving the night count, and totalling airfare plus lodging into one all-in per-person budget with filters like free_cancellation and review_score_8. Use when the user asks to plan a trip, vacation, holiday, honeymoon, city break, getaway or long weekend, or asks for flight and hotel together, a flight + hotel package, total trip cost, all-in cost, trip budget, or a per-person cost breakdown — "plan a trip to Rome for a week in May under $2000", "how much would 7 nights in Lisbon cost", "can I do Tokyo for under $3k", "10 days in Japan for a family of four", "JFK to FCO plus a hotel, budget €1500", "long weekend in Barcelona for two, £800 all in", or fit an itinerary inside a price cap in any currency. Returns the aligned checkin_date/checkout_date and night count, round-trip flights with total_price_as_number and buy_link, matching hotels with price, review_score and a booking link, and the combined per-person total against the cap. Flight-only date-range scanning is cheapest-dates, hotels alone hotel-search, picking where to go destination-compare, writing results up for a client travel-brief.

  • Reads credentials
  • Sends data out

14 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use flashcard-generator, inbox-triage and 1 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • JayRHaSkill

    flashcard-generator

    Turns notes, articles, or a topic into high-quality spaced-repetition flashcards that follow proven formulation principles — atomic, minimum-information, cloze deletions, and active recall — ready to import into Anki or any SRS. Use this skill when the user asks to "make flashcards", "create Anki cards", "turn these notes into flashcards", "help me memorize X", "generate study cards", or "make cloze deletions". Produces clean Q/A pairs and can export to Anki-importable CSV.

    Also in 3 projects that use this

  • MicrosoftSkill

    inbox-triage

    Use this skill whenever the user asks to clean up, triage, declutter, or archive low-value Outlook inbox mail - "clean up my inbox", "triage my mail", "declutter my inbox", "sort out newsletters", "archive old resolved threads", "/inbox-triage" - to group only newsletters, auto-notifications, past-event meeting logistics, resolved threads, and redundant duplicates into reviewable move-proposals presented for per-bucket approval before any message moves. Do not use for reading mail, drafting or sending replies, deleting mail, executing moves without explicit per-bucket approval, or acting on messages from the user's manager, direct reports, active threads, flagged mail, or labelled/HR/Legal/Finance/Security senders.

    Also in 3 projects that use this

  • AnthropicSkill

    interview-prep

    Create structured interview plans with competency-based questions and scorecards. Trigger with "interview plan for", "interview questions for", "how should we interview", "scorecard for", or when the user is preparing to interview candidates.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 13 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Check the echoed checkin_date, checkout_date and applied_filters in the response body against what you sent before you do any arithmetic on …

    line 88
  2. Distinctive:

    Rank results yourself: Never assume the first item is the cheapest, and if you split a trip into two one-ways (open-jaw), combine and rank …

    line 122
  3. Distinctive:

    If this file names another skill from this set (travel-data-api, cheapest-dates, destination-compare, trip-planner, fare-watch, …

    line 134

“Common” is measured: kept by at least 6 in 10 of the 13 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Trip Planner

Overview

This skill turns a fuzzy request like "a week in Rome at the end of June under $2000" into two live searches — one round-trip flight search and one hotel search on the same dates — and one combined number the user can act on. Auth is one header, x-rapidapi-key, read from the RAPIDAPI_KEY environment variable and sent only to the FlightPowers hosts named in this file (full reference: the travel-data-api skill, or https://flightpowers.com/docs). Prefer the hosted MCP tools when a client has them configured (ad-free: https://flights.flightpowers.com/mcp and https://hotels.flightpowers.com/mcp) — both legs are available over MCP: search_roundtrip_flights on the flights server, search_hotels on the hotels server.

The fields you will actually read: from flights, total_price_as_number, total_price, buy_link, price_range_in_relation_to_other_periods, price_insights_low, price_insights_high, and the departure_flight_arrival_description / return_flight_departure_description strings that tell you which calendar day each end of the trip lands on. From hotels, properties[].name, price, price_string, review_score, review_count, room_type, location, and link, plus the echoed checkin_date, checkout_date, applied_filters, and budget_per_night.

This is an independent API that returns publicly available flight and hotel pricing; it is not affiliated with, endorsed by, or sponsored by Google or Booking.com.

Endpoints

LegEndpoint / MCP tool
Round-trip flightsPOST /api/google_flights/roundtrip/v1 on google-flights-live-api.p.rapidapi.com
Round-trip flights (MCP)search_roundtrip_flights — takes departure_date_from/departure_date_to, a list of destinations, and nights
One-way legs (open-jaw)POST /api/google_flights/oneway/v1 on google-flights-live-api.p.rapidapi.com
One-way legs (MCP)search_oneway_flights — takes departure_date_from/departure_date_to and a list of destinations
Hotels for the stayPOST /search on booking-live-api.p.rapidapi.com
Hotels for the stay (MCP)search_hotels on the hotels server
One named hotelPOST /hotel_by_name on booking-live-api.p.rapidapi.com

Workflow

  1. Resolve the request into dates and airports before calling anything. Convert the destination to an IATA airport code (Rome → FCO). Turn "a week in May" into a concrete departure_date and return_date. Ask for the origin airport if the user did not give one — never guess it, a wrong origin makes every number downstream wrong.
  2. Derive nights from the dates, not from the word "week". nights = checkout_date − checkin_date in days. 2027-06-26 → 2027-07-03 is 7 nights (8 days on the ground). A "week-long trip" is 7 nights only if the user means 7 sleeps; if they said "7 days", confirm whether that is 6 or 7 nights before you price anything.
  3. State the cost before you spend it. Every search is a billed request against the user's own key. Say "this is 2 calls — one flight search, one hotel search" up front. If you are comparing 3 date windows × 2 cities, that is 6 flight calls plus 6 hotel calls; get agreement first.
  4. Search flights first. The flight is the larger and least flexible number, and it sets the hotel envelope. Pass max_price as the flight ceiling for the WHOLE party (it is compared with the total for every passenger, not per person), currency, passengers[] (1 adult, 2 child, 3 infant on lap, 4 infant in seat), and sort_type: "Price".
  5. Align checkin_date / checkout_date to the flight, not to the request. Read departure_flight_arrival_description (e.g. "8:50 AM on Mon, Jun 1"). On a red-eye or long-haul the outbound lands the day after departure_date — in that case checkin_date is the arrival day, not the departure day. Likewise if return_flight_departure_description is an early-morning slot, the traveller still needs the previous night, so checkout_date stays the return day. Recompute nights after this adjustment.
  6. Compute the hotel envelope, then search hotels. With a per-person budget B, the flight total for the whole party F = total_price_as_number (a round trip for two adults is priced for both), N nights and A adults sharing one room: room_budget = (B × A) − F and budget_per_night = room_budget ÷ N. Worked example: B = 2000, A = 2, F = 1800 (the flight ceiling below), N = 7 → room_budget 2200 → about 314 a night. Send that as budget_per_night, with filters: ["free_cancellation", "review_score_8"] so the shortlist is refundable and well-reviewed.
  7. Combine and present. Per-person total = (F + hotel_stay_cost) ÷ A, where A is every traveller the budget is per head of (children included if their fares are in F). Show the gap to budget, the price_range_in_relation_to_other_periods verdict against price_insights_low/price_insights_high, both live links, and the timestamp of the search.

Search the Round-Trip Flight

: "${RAPIDAPI_KEY:?set RAPIDAPI_KEY to your FlightPowers key}"

curl -sS -X POST \
  "https://google-flights-live-api.p.rapidapi.com/api/google_flights/roundtrip/v1" \
  -H "x-rapidapi-key: ${RAPIDAPI_KEY}" \
  -H "x-rapidapi-host: google-flights-live-api.p.rapidapi.com" \
  -H "Content-Type: application/json" \
  -d '{
    "from_airport": "JFK",
    "to_airport": "FCO",
    "departure_date": "2027-06-26",
    "return_date": "2027-07-03",
    "passengers": [1, 1],
    "seat_type": 1,
    "currency": "USD",
    "max_price": 1800,
    "sort_type": "Price",
    "max_departure_stops": 1,
    "max_return_stops": 1,
    "limit": 10
  }'

Search Hotels for Exactly Those Nights

checkin_date and checkout_date are the flight dates from the call above. The date span is the night count — 2027-06-26 to 2027-07-03 is 7 nights.

: "${RAPIDAPI_KEY:?set RAPIDAPI_KEY to your FlightPowers key}"

curl -sS -X POST \
  "https://booking-live-api.p.rapidapi.com/search" \
  -H "x-rapidapi-key: ${RAPIDAPI_KEY}" \
  -H "x-rapidapi-host: booking-live-api.p.rapidapi.com" \
  -H "Content-Type: application/json" \
  -d '{
    "destination": "Rome",
    "checkin_date": "2027-06-26",
    "checkout_date": "2027-07-03",
    "adults": 2,
    "children": 0,
    "currency": "USD",
    "budget_per_night": 310,
    "filters": ["free_cancellation", "review_score_8"]
  }'

Check the echoed checkin_date, checkout_date and applied_filters in the response body against what you sent before you do any arithmetic on properties[].

Repeat Variant — Several Windows or Several Cities

There is no bulk trip endpoint. Comparing windows means repeating the pair of calls, and each iteration is a separate billed request — quote the total first.

: "${RAPIDAPI_KEY:?set RAPIDAPI_KEY to your FlightPowers key}"

for WINDOW in "2027-06-19|2027-06-26" "2027-06-26|2027-07-03" "2027-07-03|2027-07-10"; do
  DEP="${WINDOW%%|*}"
  RET="${WINDOW##*|}"
  echo "=== depart ${DEP} / return ${RET} ==="
  curl -sS -X POST \
    "https://google-flights-live-api.p.rapidapi.com/api/google_flights/roundtrip/v1" \
    -H "x-rapidapi-key: ${RAPIDAPI_KEY}" \
    -H "x-rapidapi-host: google-flights-live-api.p.rapidapi.com" \
    -H "Content-Type: application/json" \
    -d "{\"from_airport\":\"JFK\",\"to_airport\":\"FCO\",\"departure_date\":\"${DEP}\",\"return_date\":\"${RET}\",\"passengers\":[1,1],\"currency\":\"USD\",\"sort_type\":\"Price\",\"limit\":5}"
  curl -sS -X POST \
    "https://booking-live-api.p.rapidapi.com/search" \
    -H "x-rapidapi-key: ${RAPIDAPI_KEY}" \
    -H "x-rapidapi-host: booking-live-api.p.rapidapi.com" \
    -H "Content-Type: application/json" \
    -d "{\"destination\":\"Rome\",\"checkin_date\":\"${DEP}\",\"checkout_date\":\"${RET}\",\"adults\":2,\"currency\":\"USD\",\"filters\":[\"free_cancellation\",\"review_score_8\"]}"
done

When an MCP client is configured, do the flight half of this in one search_roundtrip_flights call: pass departure_date_from: "2027-06-19", departure_date_to: "2027-07-03" and nights: 7. The range bounds the departure date, not the return, so the last window in the loop above departs 2027-07-03 and returns 2027-07-10 — setting departure_date_to to the return date would silently price an extra window. One user intent is one tool call — never one call per date. That call prices every departure day in the range (15 searches here, not 3), so quote 15 before running it; on the free sign-in allowance it is more than the day's 10 and is refused, so use the 3-window REST loop or set max_searches. The nights value you pass there is the same night count you feed to checkin_date/checkout_date on the hotel call.

Common Pitfalls

  • An empty flight result is [] with HTTP 200, not an error. It means "no flights on this route for these dates" only when X-Search-Status is ok or empty — then report it plainly and offer adjacent dates or a nearby airport, without retrying the identical call. On degraded the search did not complete: do not tell the user the trip is impossible, and do not price a trip around a flight leg you never actually searched.
  • A sold-out hotel is available: false, not a failure. On /hotel_by_name a sold-out property returns available: false with null price fields — that is a valid, correct answer, so say "not available for those nights" rather than reporting an error. On /search the equivalent is an empty properties array; loosen budget_per_night or drop a filter and say that you did.
  • Rank results yourself. Never assume the first item is the cheapest, and if you split a trip into two one-ways (open-jaw), combine and rank the two halves yourself.
  • On a long-haul or thin route, an empty result is often the search failing, not genuine unavailability. That is exactly what X-Search-Status: degraded tells you: retry on that signal, once and announced. Leave use_fallback unset: unset already lets a search whose own retries failed escalate once to a second data source, while true forces that source on every attempt, which is much slower and can time out.
  • Every combination is a billed request. Three date windows across two cities is twelve calls, not one. Price the plan out loud before running it and let the user cut the matrix down.
  • Never reuse or cache a fare. Fares and room rates go stale in minutes. Do not carry a price forward from earlier in the conversation, from an example, or from a previous window — re-search and stamp the result with the time it was fetched.
  • /search returns no nights field. Derive nights from the dates you sent, and read price_string before you multiply anything. Do not silently turn a per-night price into a stay total; if the basis is ambiguous, quote price_string verbatim next to your own arithmetic and label which is which.
  • Keep the passenger set consistent across both legs. passengers[] on the flight call and adults/children on the hotel call must describe the same travellers, or the per-person total is meaningless. State the split you used (e.g. "2 adults, one shared room").

Output Standards

  • Lead with the verdict against the budget — "7 nights in Rome, 2027-06-26 to 2027-07-03: about $X per person, $Y under your $2000" — then the flight line, the hotel line, and the arithmetic that joins them, so the user can check it. Always show the derived night count and the exact checkin_date/checkout_date you searched.
  • Stamp every total with when it was fetched, quote the price_range_in_relation_to_other_periods verdict alongside price_insights_low/price_insights_high so "$209 is typical for this route, don't rush" is available instead of a bare number, and hand over the live buy_link and hotel link for booking. Never present a fare or rate as booked or guaranteed — prices are live search results and can change before the user finishes clicking.

If this file names another skill from this set (travel-data-api, cheapest-dates, destination-compare, trip-planner, fare-watch, hotel-search, rate-parity-monitor, travel-brief) that is not installed, do that part of the job with the endpoints in this file.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.